Content uploaded by Roger Enriquez
Author content
All content in this area was uploaded by Roger Enriquez on Feb 10, 2017
Content may be subject to copyright.
The Effects of Social Disorganization:
A Hierarchical Analysis
of Perceived Incivilities
in a Latino Community
Jeffrey Michael Cancino
Sean Patrick Varano
Joseph A. Schafer
Roger Enriquez
ABSTRACT. Guided by the theoretical framework of quality of life and
social disorganization, this study combined data from three independent
sources (4,469 community surveys, Census Bureau, and police crime
records) to simultaneously examine the influence of contextual character-
istics (concentrated disadvantage, social isolation, and violent crime) on
residents’ perceived incivilities across 10 city council districts in San
Antonio, Texas, net of citizen-level covariates. Several findings emerged
from the Poisson hierarchical analysis. At the citizen-level, Latino, age,
Jeffrey Michael Cancino is affiliated with Department of Criminal Justice, Texas
State University-San Marcos, 601 University Drive, San Marcos, TX 78666.
Sean Patrick Varano is affiliated with College of Criminal Justice, Northeastern
University, 204 Churchill Hall, Boston, MA 02115.
Joseph A. Schafer is affiliated with Center for the Study of Crime, Southern Illinois
University, Mailcode 4504, Carbondale, IL 62901.
Roger Enriquez is affiliated with Department of Criminal Justice, University of
Texas-San Antonio, 501 West Durango Boulevard, San Antonio, TX 78207.
Address correspondence to: Jeffrey Michael Cancino, Department of Criminal
Justice, Texas State University, 601 University Drive, San Marcos, TX 78666 (E-mail:
jc68@txstate.edu).
Journal of Ethnicity in Criminal Justice, Vol. 5(1) 2007
http://jecj.haworthpress.com
©2007 by The Haworth Press, Inc. All rights reserved.
doi:10.1300/J222v05n01_01 1
homeowner, perceived safety, and quality of life rating were significantly
related to perceived incivility. At the council district level, all the contextual
variables were positively and significantly associated with the outcome.
Implications from these findings are considered. doi:10.1300/J222v05n01_01
[Article copies available for a fee from The Haworth Document Delivery Ser-
vice: 1-800-HAWORTH. E-mail address: <docdelivery@haworthpress.com>
Website: <http://www.HaworthPress.com> © 2007 by The Haworth Press, Inc.
All rights reserved.]
KEYWORDS. Social disorganization, incivilities, Latinos, council
districts
While criminal victimization is a serious matter, citizens’ concern
over their well-being goes beyond crime (Taylor & Shumaker, 1990).
For example, Garofalo and Laub (1978) established that subjective
quality of life evaluations affect a much wider audience than actual
crime. Among the many quality of life variables to choose from (e.g.,
fear), this study focuses on incivilities (also known as disorder). There
is a general consensus within the criminal justice community regarding
what constitutes incivilities. Social incivilities are commonly character-
ized by public drinking, drunkenness, unsupervised teens, noisy neigh-
bors, gangs, prostitutes, and the like; whereas, physical incivilities
reflect abandoned housing, graffiti, vacant trash-filled lots, and aban-
doned cars. Over the years, social scientists have developed two dis-
tinct, yet theoretically similar, lines of inquiry when studying social and
physical incivilities.
On the one hand, scholars have investigated how perceived citizen
incivilities stimulate concern over subjective quality of life outcomes
such as safety, fear, and risk of victimization (e.g., Ferraro, 1994;
Taylor, 2001; Wilson & Kelling, 1982); on the other, researchers have
endorsed a contextual perspective by relying on social disorganization
theory to evaluate how adverse structural constraints of the local com-
munity (e.g., poverty, residential instability, ethnic heterogeneity) in-
fluence citizens’ perceived incivilities (Aneshensel & Sucoff, 1996;
Reisig & Cancino, 2004; Taylor & Covington, 1993). Sampson and
Raudenbush (1999, p. 626) empirically concluded that disorder is a
product of weakened social controls and structural antecedents associ-
ated with social disorganization. Whether modeling incivilities as an in-
dependent or dependent variable, these studies share similar theoretical
2 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
explanations. For instance, a citizen that reports incivilities may express
higher levels of fear and risk of victimization; which in turn, mayunder
-
mine his/her ability to participate in developing informal social control
(Skogan, 1990, p. 47). Likewise, citizens that perceive higher levels
of incivilities may infer that such neighborhood is becoming socially
disorganized due to a breakdown of informal social control.
The present research contributes to the emerging body of literature
that (1) seeks to better understand how the context influences citizen
perceived incivilities, and (2) extends the study of incivilities to Latino
communities. Studying incivilities among Latinos are important for
four reasons. First, existing scholarship reports mixed evidence in terms
of whether Latinos report higher (or lower) levels of incivility com-
pared to Whites and Blacks. Second, contextual studies tend to impli-
cate the structure as a major source of perceived incivility. Third, Latinos
are likely to reside in communities characterized by social disorganiza-
tion. Fourth, incivilities represent a weakening of the community’s in-
formal regulatory capacity (Sampson & Radenbush, 1999; Wilson &
Kelling, 1982). By enhancing our knowledge about incivilities, we can
better understand the proximal and distal sources that shape citizens’
perceptions. Perceived incivilities are defined, here, as survey respon-
dents’ personal judgments and state of their residential surroundings
with respect to signs of social and physical decay. Reiss (1973, p. 392)
argues that “measuring the quality of life in a community or society is
no simple matter since what is at stake are human values, judgments,
and subjective perceptions of social reality.”
The current research is guided by the conceptual and theoretical
framework of quality of life and social disorganization. This study
combined data from three independent sources (community surveys,
Census Bureau, and police crime records) to simultaneously examine,
in a multilevel fashion, the influence of contextual characteristics (con-
centrated disadvantage, social isolation, and violent crime) on resi-
dents’ perceived incivility across city council districts in San Antonio,
Texas, net of citizen-level covariates. The current Poisson hierarchical
strategy and research setting provided a rare opportunity to study both
citizen- and contextual-level attributes of a predominately Mexican-
American Latino community. Social scientists are encouraged to con-
sider Latinos in ways that move beyond the White-Black quality of
life dichotomy.
Cancino et al. 3
LITERATURE REVIEW
While incivilities have been studied at length,1there remains a sub-
stantial deficit in knowledge when discussing Latinos and incivilities.
The dilemma, as demonstrated by the review later, is that individual-level
studies seem to include Latinos yet seldom specify perceived incivility as
the outcome. In contrast, studies that consider both individual- and con-
textual-level influences more frequently specify perceived incivility as
the outcome but rarely include Latinos. The goal is to adjudicate these
concerns, and in the process, cast light on the Latino-incivilities connec-
tion. The literature review was divided in two subsections. The review
begins with quality of life individual-level studies, followed by an over-
view of multilevel studies that also consider contextual characteristics.
Individual-Level Studies
Researchers have examined a broad range of quality of life outcomes,
including perceived crime, fear of crime, risk of victimization, and safety,
which are grounded in social psychology and reflect emotional, affective,
and cognitive reactions (Ferraro, 1994; Ferraro & LaGrange, 1987, p. 78;
Warr, 2000). Among the various quality of life variables to study, it ap-
pears that scholars favor outcomes related to perceived fear of crime and
risk of victimization when studying Latino populations. Despite the ne-
glect of incivilities, the subsequent studies are informative since they pro-
vide a starting point regarding (1) what Latinos perceive as affecting their
quality of life and (2) how evaluations compare to their racial/ethnic
counterparts. In a series of recent publications, Lane and Meeker used a
1997 telephone survey instrument administered to 1,000 Orange County,
CA, residents to examine how perceived social disorganization condi-
tions (i.e., community minority diversity, disorder, decline) influenced
quality of life assessments. Lane and Meeker (2003) found that perceived
increase in Latino populations (i.e., community minority diversity) was
significantly related to fear of gang crime. Residents also reported higher
levels of fear in association with a perceived weakening of local social
ties (i.e., community decline), and a combined community decline and
disorder (e.g., graffiti, trash, and gangs) model. Furthermore, Latinos
were more concerned about disorder, but Whites were more concerned
about the breakdown of local social ties.
4 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Using the same data, Lane and Meeker (2004; see also Chiricos et al.,
2001) also employed ANOVA to investigate whether Whites, Latinos,
or Vietnamese were more concerned with perceived community disor-
der, diversity, risk of victimization, and fear of gang crime. Vietnamese
reported significantly higher levels of disorder, decline, risk, and fear,
followed by Latinos and Whites. Although Vietnamese and Latinos re-
sided in socially disorganized areas, the latter group perceived Latino
gangs as part and parcel of the community context, which in turn, yielded
less reaction to risk and fear. In their latest study, Lane and Meeker
(2005) found that for Latinos, concern about community diversity in-
creased perceptions of disorder and consequently gang fear; but per-
ceived community decline was unrelated to gang fear. For Whites, as
perceived diversity concerns increased so did perceived disorder, which
enhanced concern about decline and gang fear.
The aforementioned studies suggest that citizens’ subjective state of
their neighborhood affect how they evaluate their quality of life. What
makes these “quality of life” studies particularly compelling is the in-
clusion of Latino populations. For example, each demonstrated a Latino
sample of 18, 28, and 30 percent respectively. There is one overarching
limitation. The research failed to adjust for contextual characteristics.
Models estimating both individual and contextual predictors may pro-
vide a more informed understanding regarding quality of life. In fact,
social ecology theorists argue that quality of life perceptions extend
beyond an individual’s subjective surroundings, and are more closely
associated with objective social disorganization conditions.
Contextual and Advanced Multilevel Studies
Social disorganization posits that communities characterized by pov-
erty, residential instability, and ethnic heterogeneity lead to a break-
down in informal social control, which in turn, influences rates of crime
and delinquency (Shaw & McKay, 1942). Over the years, theoretical
adjustments have been made, especially with regards to poverty (see
Duncan & Aber, 1997), and current research demonstrates that such
theory has been used to explain quality of life outcomes (Sampson &
Raudenbush, 1999; Taylor & Covington, 1993). The premise is that
citizen evaluations about where they live reflect multiple social disorga-
nization structural conditions. Still working in the social ecology tradi-
tion, researchers have also shifted attention to assessing the impact of
segregation on violent crime among African-American communities
Cancino et al. 5
(e.g., Lee & Ousey, 2005; Peterson & Krivo, 1993). According to Wilson
(1987) and Massey and Denton (1993), segregated minority communities
lack access to mainstream ethos by (1) limiting exposure to individuals
who maintain stable families and participate in community issues and
(2) eroding the economic opportunity structures of residents by restrict-
ing access to job networks. It is important to underscore that the concept
of segregation has multiple and empirically distinct geographic forms
(e.g., social isolation) that may influence crime-related outcomes in dif-
ferent ways (Massey & Denton, 1988; Shihadeh & Flynn, 1996).
The current study focused on social isolation as opposed to the more
commonly used residential segregation because the former has been
credited with reducing the level of informal social control in African-
American communities since the consequences often lead to multiple
economic, political, cultural, and social disadvantages (Shihadeh &
Flynn, 1996, p. 1329). Wilson (1987, p. 60) defined social isolation as
“a lack of contact or of sustained interaction with individuals and insti-
tutions that represent mainstream society.” The assumption, here, is that
social isolation may operate similarly in predominant Latino communi-
ties. Burton (2004), for example, found that social isolation (but not res-
idential segregation) significantly predicted Latino homicides across
113 cities. Hence, it is possible that citizens residing in communities
characterized as being socially isolated from mainstream society may
report higher levels of incivility.
Multilevel statistical modeling (e.g., HLM) has been used more
recently to understand the complex interactions between individuals
and their contextual surroundings. Lee and Ulmer (2000), for exam-
ple, examined simultaneous effects of several individual- and contex-
tual-level variables on perceived incivility (and three other quality of
life outcomes) among Korean-Americans in Chicago. They found that
community crime rate, population density, citizen self reported in-
come, and other citizen-level predictors were positively associated
with the outcome. Reisig and Cancino (2004) used a hierarchical
Poisson regression analysis to test social disorganization effects in
nonmetropolitan Michigan communities and discovered that respon-
dents living in economically disadvantaged areas were significantly
more likely to report higher levels of perceived incivility. In another
study, Reisig and Parks (2004) reported that police community collab-
oration mediated the effect of neighborhood concentrated disadvan-
tage on perceived incivility. Thus far, these multilevel studies fail to
include a sufficient Latino sample size. However, there is a handful of
6 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
scholarship that has included a moderate to large Latino sample and
considers incivilities as the outcome.
Skogan and Steiner’s (2004) sample of 1,007 Latinos nested in 157
police beats in Chicago showed that concentration of Latinos and pov-
erty were positively related to perceived physical and social disorder.
They also reported that Latinos experienced a significant decrease in
quality of life over the two-year study period; a decrease not experi-
enced by Whites and African-Americans. Aneshensel and Sucoff (1996)
examined the effect of poverty, segregation, and perceived disorder on
perceived mental health outcomes among 877 adolescents (approxi-
mately 50% of the sample was Latinos) in California. In general, youth
residing in impoverished and minority segregated communities re-
ported higher levels of perceived disorder, which in turn, influenced
self-reported symptoms of mental well-being. These results are consis-
tent with Massey and Denton’s (1993) description of a self-sustaining
cycle of social disorder and social withdrawal stemming from social
isolation. Aneshensel and Succoff’s (1996) findings reinforce the notion
that neighborhood context influences perceived incivilities in ways that
lead to a downward spiral of social, emotional, and physical decay. An-
other intriguing finding was that Latino adolescents reported less de-
pression and exhibited less conduct disorder in poor neighborhoods
with dense concentrations of Latinos. One explanation for this discov-
ery is that dense Latino populations may serve as a support and deterrent
mechanism that buffers the impact of adverse contextual effects on qual-
ity of life.
Arguably, the most comprehensive study to date is Sampson and
Raudenbush’s (2004) multilevel examination of 3,585 (33% Latino sam-
ple) citizens living in 478 Chicago census block groups. Building on the
racial stigma framework which is part of the larger segregation concept
(Loury, 2002; Massey & Denton, 1993), the authors claim that racial
composition and poverty are likely to yield elevated levels of perceived
disorder after controlling for observed signs of disorder and violent
crime. Regarding the former concept, the make-up of a community’s
White, Black, and Latino composition carries powerful cultural stereo-
types and stigmas that implicitly bias citizen notions of vulnerability,
crime, and quality of life concerns (Banji, 2002; Quillian & Pager, 2001).
The connection between racial composition and disorder can promote
segregation. For example, White (or Latino) citizens may implicitly for-
mulate and socially construct a bias that Blacks are associated with
a rise in crime and perceived disorder. As a result of such stereotype
against Blacks, Whites may then choose to leave these areas; which in
Cancino et al. 7
turn, creates segregated communities (Sampson & Raudenbush, 2004,
p. 322; see also Charles, 2003). Sampson and Raudenbush (2004) found
that while observed disorder and violent crime predicted perceived dis-
order, the concentration of racial groups and poverty were more impor-
tant predictors for all races. In addition, “as percent Black in the block
group increases (but not percent Latino), Latino respondents tend to
perceive significantly more disorder than do Whites” (pp. 332-333).
In summary, in these multilevel studies, community-level predic-
tors explained the greatest proportion of variance when compared to
individual-level predictors. The literature also provided considerable
evidence that personal quality of life assessments are a function of
objective ecological characteristics, such as concentrated disadvan-
tage, segregation, and crime. Thus, the limited research that includes
Latino populations seems to implicate the structure as a major source
of perceived incivility. Such a notion remains open to empirical inves-
tigation, here.
RESEARCH OBJECTIVE
The research objective was to simultaneously examine the influence
of contextual characteristics on citizens’ perceived incivility across city
council districts in San Antonio, Texas, net of citizen-level covariates.
To accomplish this objective, three multivariate models were estimated
which reflected different levels of explanation (i.e., citizen vs. council
district). Since the models represented different levels of explanation, a
multilevel strategy was endorsed. Overall, the research aim was to de-
termine whether an important source of variation lies in the differential
ecological structure of council districts.
DATA AND METHODS
Based on socio-demographic characteristics (see Census 2000 Sum-
mary File 1), San Antonio is a unique city that is particularly attractive
for advancing Latino research. Three independent data sources were
culled to construct the data file: community surveys, the Census Bureau,
and San Antonio Police Department (SAPD). The following subsec-
tions highlight these data sources, along with the collection and mea-
surement procedures.
8 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Community Survey Data
During the fall semester of 1998, the Hispanic Research Center at the
University of Texas-San Antonio conducted a telephone survey ques-
tionnaire of San Antonio citizens’ perceptions regarding local political
and quality of life issues. Relying on telephone numbers purchased
from an independent company, a random probability sample of 8,000
adults (i.e., 18 years or older) were surveyed. Using a Computer Aided
Telephone Interviewing (CATI) system, administration of the surveys–
conducted in English and Spanish–were carried out by trained under-
graduate and graduate students under the supervision of faculty. From
the original sample of 8,000 adults, there were 5,013 residents that par-
ticipated.2Due to missing survey-item information, the sample size of
5,013 was further reduced by 544 questionnaires. The final sample con-
sisted of 4,469 complete and usable surveys. Overall, the response rate
was 71 percent. Consistent with previous research that employed the
CATI system (Lane & Meeker, 2004), the response rate was calculated
by taking the number of completes / [number of completes !refusals].
Official Census Bureau and Police Crime Data
The second source of data was the 2000 decennial census. Census
tract data was aggregated to the council district boundaries across the
city. Following the criteria established by urban sociologists that used
physical landmarks, census data, and first-hand knowledge to construct
geographic aggregates (Sampson et al., 1997, p. 919; also see Small &
Newman, 2001, p. 31), council districts represented the contextual-level
unit of analysis. Sampson and Groves (1989) have used similar units of
analysis, such as wards and polling districts. Nonetheless, council dis-
tricts are viewed as independent and meaningful. Given the theoretical
explanation that perceived incivilities signals a breakdown of informal
social control associated with social disorganization, council districts are
meaningful because citizens are more likely to have direct contact with
their elected council representative. The final source of data was secured
from the SAPD. This department provided official violent crime (e.g.,
homicide, robbery, aggravated assault, and forcible rape) incident records
for 1998. The physical address where the incident occurred was geo-
coded to its respective council district.
Overall, the analysis file used in this study consisted of 4,469 citi-
zens nested within 10 council districts. The number of survey respon-
dents in each council district ranged from 403 to 654 (average = 446;
Cancino et al. 9
median = 434). Researchers have achieved reliable results using at least
10 cases within 10 aggregates when performing multilevel analyses
(Mok & Flynn, 1998; Welsh, Greene, & Jenkins, 1999). Additional dia-
logue on this issue is warranted and detailed in the discussion section.
Dependent Variable
The dependent variable, perceived incivility, was a count variable
that asked survey respondents about the number of serious physical and
social disorder problems in their local communities. Perceived incivil-
ity included the following items: abandoned buildings, vacant lots, and
gangs. Each incivility item was dummy coded (1 = Yes, 0 = No). Items
were summed to create a three-item additive scale of physical and social
disorder. The scale exhibited a satisfactory level of internal consistency
(Cronbach’s "=.70).Previousstudieshaveusedsimilarincivilityitems
and count variable measures (Reisig & Cancino, 2004; Reisig & Parks,
2004; Sampson & Raudenbush, 1999).
Citizen- and Council District-Level Independent Variables
Citizen-Level Variables. Ethnic and racial differences in perceived
incivility were examined by estimating two dummy variables: Latino
(1 = Latino, 0 = Other) and non-Latino minority (1 = Non-Latino minor-
ity, 0 = Other). It was hypothesized that Latino and non-Latino minority
would be positively associated with perceived incivility. Prior research
suggests that various socio-demographic characteristics may also influ-
ence the outcome variable (Skogan & Maxfield, 1981, p. 75). Thus, ad-
ditional citizen-level predictors were included in the analysis to control
for survey response bias and spuriousness: male (1 = Male, 0 = Female),
age (respondent’s age in years), homeowner (1 = Owner, 0 = Other),
and education (1 = Not a high school graduate, 2 = High school gradu-
ate, 3 = Some college, 4 = College graduate, 5 = Some postgraduate
work). The present study also included two variables consistent with
what Skogan (1990) described as citizens’ subjective quality of life
judgments of their local surroundings. Perceived safety was a single
survey item (“Thinking about where you live, are you concerned about
your safety?”) that included a dummy coded response set (1 = Yes, 0 =
Other). Using a more global measure, quality of life rating was oper-
ationalized as a single survey item (“Thinking about your neighbor-
hood, how would you rate your quality of life?”) that featured four
response categories (1 = Poor, 2 = Fair, 3 = Good, 4 = Excellent). It was
hypothesized that perceived safety would be positively associated with
10 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
perceived incivility; whereas quality of life rating would be inversely
associated with the outcome.
Council District-Level Variables. To capture the contextual arraign-
ment of council districts, three variables were estimated. Guided by
the social disorganization tradition showing that structural characteris-
tics effect levels of disorder (Sampson & Raudenbush, 1999, p. 24;
Skogan, 1990, p. 60), concentrated disadvantage was operationalized
as a weighted factor regression score (eigenvalue = 2.75, factor loadings #
80) that included the following 2000 Census items: percent poverty,
percent public assistance, percent unemployment, percent female-headed
household with children, and percent minority.
Among the various ways that minorities can be segregated from
Whites (see Massey & Denton, 1988 for a comprehensive discussion),
segregation is conceptualized as social isolation. Social isolation is de-
fined as the degree of potential contact between minorities and Whites.
There exist two reasons for selecting social isolation over the more
commonly used residential segregation that measures unevenness. First,
according to Shihadeh and Flynn (1996, pp. 1137-1138), social isola-
tion is better able to capture the manner in which minorities and Whites
are distributed across geographic aggregates. Second, Burton (2004)
posits that social isolation is a more accurate indicator of deteriorating
social conditions than residential segregation. Thus, social isolation
was operationalized consistent with Lieberson’s (1981) interaction in-
dex that gauged the probability, mP*w, that a randomly drawn minority
in a council district has contact with a White individual:
mwi
iii
Pmwt
*(/)=Â
where mirefers to the proportion of all minorities in the city that are
located in council district i, and wiand tiare the number of Whites and
total population in council district i. Values range from 0 to 1.0. The
score of 0 indicates complete social isolation of minorities from Whites.
Since San Antonio’s minority population is 58 percent Latino and 6 per-
cent African American, it was decided to combine both racial and ethnic
groups in the index.
Similar to prior research that encourages scholars to control for crime
measures when studying incivilities (Reisig & Parks, 2004; Sampson &
Raudenbush, 1999; Sampson & Raudenbush, 2004), a crime variable was
constructed. Violent crime index was measured as a weighted factor score
that included official police recorded crime per 100,000 residents for the
Cancino et al. 11
six months prior to conducting the survey and aggregated citizen reported
perceptions of violent crime (percent reporting that crime was “a prob-
lem” or “not a problem” where they lived) (eigenvalue = 1.30, factor
loadings #.80). Researchers have estimated comparable crime measures
(Reisig & Cancino, 2004; p. 21; Sampson & Jeglum-Bartusch, 1998,
p. 798). It was hypothesized that concentrated disadvantage, social isola-
tion, and violent crime index would be positively associated with per-
ceived incivility. Table 1 shows descriptive statistics for citizen- and
council district-level variables.
ANALYSIS AND FINDINGS
Bivariate Associations and Model Diagnostic
Table 2 presents bivariate associations (Pearson’s r) between citizen-
level variables. The magnitudes of correlations were rather weak which
initially indicated no multicollinearity problems.3As expected, Latinos
12 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
TABLE 1. Descriptive Statisticsa
Variable Mean SD Minimum Maximum
Citizen-level
Perceived incivility .78 .98 .00 3.00
Latino .49 .50 .00 1.00
Non-Latino minority .06 .23 .00 1.00
Male .38 .49 .00 1.00
Age 41.20 9.30 18.00 86.00
Homeowner .69 .46 .00 1.00
Education 2.74 1.24 1.00 5.00
Perceived safety .23 .42 .00 1.00
Quality of life rating 2.90 .80 1.00 4.00
Council district-level
Concentrated disadvantageb.00 1.00 $1.44 1.72
Social isolation .36 .28 .03 .93
Violent crime indexb.00 1.00 $1.66 1.61
aTotal sample size is 4,469 citizens and 10 council districts.
bWeighted factor score.
TABLE 2. Bivariate Associations
1234 5 6789
Bivariate relationships between citizen-level variables used in multivariate analysis (
N
= 4,469)
Latino 1.0 $.24** $.03* $.21** $.05** $.32** .08** $.17** .14**
Non-Latino minority 1.0 $.01 .17** .01 $.15** .22** .02 .26**
Male 1.0 $.07** $.04** .09** $.02 $.01 $.03
Age 1.0 .31** $.05** $.02 .02 .02
Homeowner 1.0 .09** $.02 .09** .02
Education 1.0 $.09** .22** $.15**
Perceived safety 1.0 $.27** .50**
Quality of life rating 1.0 $.34**
Perceived incivility 1.0
Bivariate relationships between council district-level variables used in multivariate analysis (
N
=10)
Concentrated disadvantagea1.0 .41 .54 .72**
Social isolation 1.0 .19 .34*
Violent crime indexa1.0 .63*
Mean perceived incivility 1.0
aWeighted factor score.
*
p
%.05; **
p
%.01.
13
reported significantly higher levels of perceived safety (r= .08) and
incivility (r=.14).Latinosalsoreportedasignificantlylowerqualityof
life rating (r=$.17). In contrast, non-Latino minorities felt significantly
less safe (r= .22). The observed significant relationship between non-
Latino minority and perceived incivility (r=.26)wasdoublethatofLa
-
tino and perceived incivility (r=.14).Otherassociationsworthhighlight
-
ing were the healthy quality of life relationships. For example, the
strongest magnitudes were observed between perceived safety and per-
ceived incivility (r=.50)andqualityofliferatingandperceivedincivility
(r=$.34). These findings, in particular, reinforce the need to include
additional measures of quality of life to guard against spuriousness.
Table 2 also presents the bivariate associations between council dis-
trict-level variables. The contextual predictors yielded much larger cor-
relations. For example, concentrated disadvantage (r= .72), social
isolation (r= .34), and violent crime index (r= .63) were significantly
and positively associated with aggregated council-district incivilities
(i.e., mean perceived incivility). While the magnitudes were fairly strong,
prior studies have found similar associations for concentrated disadvan-
tage and crime in relation to perceived incivility (Reisig & Cancino,
2004, p. 23; Reisig & Parks, 2004, p. 155). According to the diagnostics,
multicollinearity is not a problem and unbiased council district effects
can be estimated.4
Hierarchical Regression Models
A three-step modeling strategy was enlisted using HLM Version 6.0.
First, an unconditional one-way ANOVA with random effects model
(no predictors at either level were specified) for perceived incivility was
estimated. This model is instructive, in that, it provides descriptive sta-
tistics to help determine whether (1) variation in the outcome lies within
and between council districts, (2) the sample mean computed for each
council district are reliable estimates of the true council district means
for perceived incivility, and (3) the data are suitable for estimating more
sophisticated multilevel hierarchical models (e.g., ANCOVA with ran-
dom effects and fixed effects). The findings showed that reliability esti-
mates for perceived incivility (.97) indicate that sample means are reliable
measures, and council district differences can be modeled with a high
degree of accuracy.5Reisig and Parks (2004, p. 158) found a similar re-
liability for perceived incivility.
14 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Next, the interclass correlation coefficient (ICC) was calculated
()rt/s t
00
=+
2
00 to gauge the proportion of variance in perceived inci-
vility that was between council districts. The ICC for perceived incivility
was .14; which means that 14 percent of the variance in the outcome was
between council districts. Interpreted differently, the amount of varia-
tion from citizen to citizen within council districts plus variation in mea-
surement error was 86 percent. Still drawing from the ANOVA results,
the chi-square (&2) value for the between council district variance was
used to determine whether sufficient variation existed across council
districts. The &2value (623.46, p< .000) indicated a rejection of the null
hypothesis that no significant differences in perceived incivilities ex-
isted between council districts. Overall, the findings support proceeding
with advanced hierarchical modeling techniques. Before proceeding, how-
ever, it is important to underscore that the observed citizen-level variance
('2=.84)forperceivedincivilitywaslargerthanthemean(.78)asshown
in Table 1. This suggested that over-dispersion existed. Thus, a hierarchi-
cal Poisson model designed for count data with an over-dispersed distri-
bution was selected (see, Barron, 1992, pp. 186-189; Long, 1997, p. 217;
Reisig & Cancino, 2004, p. 23; Taylor, 2001, pp. 219-225, for discussion
on count data and Poisson estimations).
The second-step was to estimate an ANCOVA random coefficient
regression model that included all citizen-level variables to ascertain
whether any of the citizen-level slopes varied significantly across coun-
cil districts. The citizen-level variables were centered around the group
means. The within-council district model was expressed as a general-
ized linear model that considered the count variable perceived incivility
as sampled from an over-dispersed Poisson distribution:
hij
ij r
=
++
log( )
PERCEIVED INCIVILITY) = X
lij
jqqij
hbb(0ij
q
Â
where (ij is the log of the event rate; lij is the event rate assuming con-
stant exposure across citizens; )0jis the intercept; Xqij is the value of
covariate qfor citizen iin council district j;)qis the partial effect of that
covariate on (ij. The citizen-specific error term, rij, is assumed to be in-
dependently and normally distributed with constant variance '2. The
random coefficient model allows the intercept to assume different val-
ues in each of the study council districts.
Cancino et al. 15
Chi-square results revealed that perceived safety (&2=42.20,p<.000),
Latino (&2=19.87,p<.05),andeducation(&2=17.63,p<.05)varied
across council districts. Since significant variation was detected, cross-
level interaction models were estimated (see, Kreft & DeLeeuw, 1998,
pp. 12-13). Here, the slope for perceived safety was modeled as a func-
tion of concentrated disadvantage, social isolation, and violent crime in-
dex. The effect of perceived safety persisted (b=.76,t-ratio = 13.78).
Concentrated disadvantage was also statistically significant (b=$.15,
t-ratio = $2.60). These results suggest that citizens’ concern regarding
safety and incivility are less pronounced in council districts characterized
by lower levels of concentrated disadvantage. Cross-level interaction
model results (not shown) for Latino and education indicated that the
slope coefficients did not vary across council districts. This latter finding
suggests that the relationship between citizen-level predictors and the
outcome were the same in all 10 council districts. Consequently, all with-
in council district slopes were specified as “fixed” across council districts
(see, Rountree et al., 1994, pp. 398-403).
In the final step, three hierarchical models were estimated to investi-
gate simultaneously citizen- and council district-level effects on the
outcome. The council district model was expressed as:
bg
0000
MEAN PERCEIVED INCIVILITY) =
j j
(gm++
Â0SSJ
W
S
)
where citizen-level intercept, )0j, is a function of council district con-
textual characteristics (e.g., concentrated disadvantage) and allowed to
vary randomly across council districts; 00 is the average value of the
outcome across council districts; 0S are the council district-level re-
gression coefficients; WSJ are the council district-level predictors; and
m0jis the unique increment to the intercept associated with council dis-
trict j(i.e., the random effect) assumed to be normally distributed with
variance *. The council district variables were centered around the
grand mean; whereas, citizen-level variables were centered around the
group mean. Table 3 presents the findings from “fixed effects” hierar-
chical Poisson multivariate models. Consistent with Long (1997), the
exponentiated coefficient was calculated in order to obtain the percent
change in Yassociated with a one-unit change in X,holdingallothervar
-
iables constant.
Model 1 included quality of life predictors, along with other citizen
covariates. This “quality of life” model was specified to represent the in-
dividual’s subjective experience of living within a city council district.
16 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Cancino et al. 17
TABLE 3. Hierarchical Poisson Regression Analysis for Perceived Incivility
Variable Model 1 Model 2 Model 3
Intercept $.34* $.45* $.46***
(.14) (.06) (.05)
Citizen-level (N = 4,469)
Latino .15* .15* .16*
(.02) (.02) (.02)
[1.16] 1.16] [1.16]
Non-Latino minority $.01 $.01 $.01
(.06) (.07) (.08)
[1.00] [1.00] [1.00]
Male $.02 $.02 $.02
(.03) (.03) (.03)
[1.02] [1.02] [1.02]
Age $.38** $.38** $.38**
(.02) (.02) (.02)
[1.46] [1.46] [1.46]
Homeowner .11*** .11** .11**
(.02) (.04) (.04)
[1.11] [1.11] [1.11]
Education .05* .05* .05
(.01) (.01) (.02)
[1.05] [1.05] [1.05]
Perceived safety .30*** .30*** .30***
(.05) (.04) (.04)
[1.35] [1.35] [1.35]
Quality of life rating $.26*** $.26*** $.26***
(.02) (.02) (.02)
[1.30] [1.30] [1.30]
Council district-level (N = 10)
Concentrated disadvantagea– .64*** .61***
(.08) (.07)
[1.91] [1.85]
Social isolation – .47** .43**
(.10) (.19)
[1.60] [1.54]
Violent crime indexa– – .42*
(.06)
[1.52]
Variance explained (percentages)
Within-council district 14 14 14
Between-council district – 72 86
Deviance statistic (
df
) – 20.27 (
7
) 16.74 (
6
)
Note:
Standard errors in parentheses and exponentiated coefficients in brackets.
aWeighted factor score.
*
p
%.05; **
p
%.01; ***
p
%.001
According to Skogan (1990), reactions to social reality are just as im-
portant as objective conditions when examining the impact of neighbor-
hood conditions on citizen’s attitudes. The quality of life model shows
that the incivility score for Latino was 1.16 times the score of Whites.
Furthermore, a one-year increase in age reduced the incivility score by
38 percent. Homeowners also reported significantly higher levels of in-
civility. As expected, the quality of life associations confirmed the di-
rectional accuracy of our hypotheses. For example, a one-unit increase
in perceived safety resulted in a 35 percent increase in perceived inci-
vility; whereas, a one-unit decrease in quality of life rating yielded a
30 percent increase in the outcome. Note that significant citizen-level
variables in Model 1 persisted across all models. Model 1 also accounted
for 14 percent of the explained variance. Overall, the findings support
the contention that citizen quality of life evaluations do influence other
quality of life outcomes; and there is a need to estimate citizen-level
covariates to guard against spuriousness.
Model 2 estimated council district social disorganization effects.
Consistent with previous research, the findings revealed that citizens re-
siding in council districts characterized by concentrated disadvantage
and social isolation were significantly more likely to report higher levels
of incivilities, independent of citizen-level covariates. Specifically, each
unit increase in concentrated disadvantage was associated with a 91 per-
cent increase in perceived incivility. Likewise, a one-unit increase in
minority segregation from Whites was related to a 60 percent increase in
the outcome. In essence, citizen quality of life evaluations are, to a much
greater extent, influenced by structural conditions. Model 2 also con-
firms our hypotheses that social disorganization is positively related to
quality of life assessments. Almost 75 percent of the explained variance
is between-council districts, as opposed to within.
Model 3 was estimated to determine whether concentrated disadvan-
tage and social isolation findings were robust enough to survive the ef-
fect of introducing violent crime as a control variable. Results indicated
that while concentrated disadvantage and social isolation remained sig-
nificant, the magnitudes of their coefficients were reduced (from .64 to
.61 and .47 to .43). Still, the two theoretical social disorganization vari-
ables were significantly and positively associated with reported incivil-
ity. Violent crime index was also significantly and positively associated
with the dependent variable. A one-unit increase in violent crime re-
sulted in a 52 percent increase in the outcome. By adding violent crime
in Model 3, a 14 percent increase in explained variance was observed
between-council districts. This indicated that there is shared explained
18 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
variance between the independent variables. Since HLM is unable to
provide an overall goodness-of-fit statistic, OLS regression deviance
statistics showed that Model 3 was significantly lower (16.74) than
Model 2 (20.27). Hence, inclusion of violent crime improves the overall
model-fit.
DISCUSSION
Before discussing the results, some important research limitations are
acknowledged. First, the data are cross-sectional in nature; which pre-
cludes definitive statements about causal linkages between the indepen-
dent measures and quality of life outcome. To better isolate causality and
simultaneity, LISREL simultaneous equation strategies that specify re-
cursive and non-recursive models are more suitable (Markowitz et al.,
2001; see also Bellair, 2000; Stucky, 2005). Second, the combined use
of three data sources introduces the problem of “shared method vari-
ance,” which can yield inflated coefficients (Bank et al., 1989). Third,
the small number of council districts (N= 10) required us to estimate
less contextual predictors than desired. By doing so, our models are
subject to potential mis-specification. A small sample also restricts
statistical power. Although Bryk and Raudenbush (1992, p. 211) artic-
ulate that a common rule of thumb is that at least 10 observations for
each predictor at the aggregate-level is preferred, they also claim that
“implausible results arising from units with small size are not a problem
because the estimation methods are robust” (p. 198). As mentioned
earlier, existing studies have used small contextual-level units of anal-
ysis (e.g., 11 aggregates, see Welsh et al., 1999). In addition, a recent
study utilized an average of 11 persons across 63 aggregates (Lee &
Ulmer, 2000). The point, here, units of analysis at level-1 and level-2
vary considerably across studies.
While aware that more precise estimates of ecological effects are
obtained by increasing the number of units at level-2, the advantages of
taking a multilevel hierarchical approach over traditional regression
strategies is superior (Blakely & Woodard, 2000, p. 370). For example,
this study was able to capitalize on the nested data structure and par-
tition citizen-from council district-level variance. Consequently, two
broad findings emerged. First, at the citizen-level, Latino, age, home-
owner, perceived safety, and quality of life rating were significantly
related to perceived incivility. Although an insignificant relationship
Cancino et al. 19
was observed for non-Latino minority and perceived incivility in the
multivariate models, the bivariate associations at the citizen-level (see
Table 2) revealed a different story. Here, the correlation between non-
Latino minority and the outcome was double (.26) that of Latinos (.14).
Yet, the non-Latino minority relationship disappeared in the multi-
variate models.
Second, at the council district-level, much larger and significant co-
efficients were observed. The findings are consistent with previous
research showing that structural constraints, such as poverty, segrega-
tion, and crime influence citizen perceived incivility above and beyond
citizen-level predictors (e.g., Aneshensel & Sucoff, 1996; Sampson &
Raudenbush, 2004; Skogan & Steiner, 2004). In terms of social isola-
tion, the descriptive statistics (see Table 1) indicated that on average
(.36) council districts were somewhat socially isolated. Moreover, the
distributions of social isolation across council districts were observed
at both tails. For example, the minimum value (.03) indicated almost
complete social isolation; whereas, the maximum value (.93) showed
almost complete interaction. As for the cross-level interaction effects,
citizens’ perceived incivility were less pronounced in council districts
characterized by lower concentrated disadvantage. Overall, the hier-
archical Poisson findings suggest that context matters in predicting qual-
ity of life across council districts.
Implications of the results indicate a need to increase quality of life
by reducing concentrated disadvantage. Researchers have argued that
concentrated disadvantage deprives citizens and communities of resources
(Land et al., 1990; Sampson & Jeglum-Bartusch, 1998). Although con-
centrated disadvantage was the strongest predictor, social isolation is
equally important given that San Antonio is a minority-majority city.
Hence, given the large number of minorities (especially Latinos), the
assumption is that there would exist a higher probability that Whites
would interact with minorities. Social isolation is also inexorably linked
with the cost of housing and living preferences. For example, based on
documented economic inequality between Whites and minorities, the
latter are less likely to afford housing in middle to upper class neighbor-
hoods occupied by the former. Research also shows that Whites avoid
living in racially heterogeneous neighborhoods (Massey & Denton,
1993). Whites that do live amongst minorities, however, report higher
levels of fear and risk of victimization (Lane & Meeker, 2004; Chiricos
et al., 2001). Clark (1992) also found that minorities preferred to live in
areas with racial compositions that were similar to their own based on
the notion that would experience less racism and stereotypes (see also
20 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Sampson & Raudenbush, 2004). While disentangling the economic
and social problems associated with concentrated disadvantage and
social isolation is difficult, the combination of these factors tend to
exacerbate community disinvestment that threatens residents’ quality
of life.
However, it is here, at the council district level, that a citizen (or
citizens) residing in a district characterized by structural constraints and
poor quality of life experiences can approach their elected city council
district representative to express their dissatisfaction. Stated differently,
while reducing concentrated disadvantage is indeed a long-term solu-
tion, citizens may receive more favorable and immediate responses
when directly contacting an elected council district official. In turn, it is
possible that favorable responses may cultivate an attitude among coun-
cil district residents that lead to additional resource deployment (e.g.,
police, employment, and housing services). Residents, however, must
exercise caution where requesting additional police service in disadvan-
taged and/or socially isolated areas as a means to reduce incivilities.
Studies suggest that depending on the context in which policing is con-
ducted, a stronger police presence may result in additional disadvan-
tages, such as more arrests and harassment (Chambliss, 1994; Hagan &
Peterson, 1995; Mosher, 2001), police perceived threat due to a large
minority population (Jacobs, 1979), and perceived minority distrust of
the police (Anderson, 1990). Nevertheless, by developing the various
kinds of services that cities offer, council district residents may collec-
tively garner support amongst each other aimed at strengthening the dis-
trict’s informal social control. The logic is that increased resources and
informal social control will improve citizen quality of life.
In conclusion, according to U.S. Census reports, the ethnic profile
of America is Latino (e.g., Guzman, 2001). Indeed, Latinos have sur-
passed African Americans as the largest minority group constituting
over 13 percent of the population (Marotta & Garcia, 2003). Future pop-
ulation estimates indicate these figures will climb due to increased im-
migration and fertility rates. “Latinos increasingly can be found in states
where there have been little or no Latinos in the past, presenting both
challenges and opportunities to service delivery systems and to policy
makers” (Marotta & Garcia, 2003, p. 14). Despite such growth, research
on Latino quality of life has remained dormant. Over the last three
years, however, an audit of the literature showed that scholars have
made great strides toward studying this ethnic group using contextual
based predictors (e.g., Holmes, 2003) and multilevel approaches. This
study adds to the small, yet engaging, body of research that examines
Cancino et al. 21
the Latino experience with the goal of bringing Latinos to the criminal
justice/criminology forefront. In the process, this scholarship seeks to
move beyond the White-Black dichotomy by considering how race/
ethnicity interacts with social conditions and quality of life.
NOTES
1. Disorder studies have a long history that date to the nineteenth century. For example,
early researchers and investigative journalists were interested in conditions associated
with health in impoverished London communities (Mayhew, 1862). Much later, Hunter
(1978) and Lewis and Maxfield (1980) coined the term incivilities to more accurately re-
flect visual and perceptual responses of disorder that threatened the civic fabric of Amer-
ican communities (see also Jacobs, 1961). Wilson and Kelling (1982) popularized the
incivilities/disorder concept and applied its fullest expression in the now classical broken
windows theory by arguing various forms of social and physical disorder directly cause
crime. Sampson and Raudenbush (1999, p. 608) have recently challenged this causal or-
dering by articulating that disorder is not a direct cause crime; instead, disorder is crime
itself. They empirically concluded that disorder and crime are both products of weakened
social controls and structural antecedents associated with social disorganization (p. 626).
2. There were 2,163 callbacks, 158 answering machines, 921 no answers, 47 fax ma-
chines, and 25 disconnected numbers, and 1,836 refusals. Note that a maximum of four
callbacks were executed.
3. To further investigate whether multicollinearity might be a concern, OLS was used
to regress perceived incivility on the individual-level variables. The diagnostic results
indicated that the tolerance statistics were sufficiently high (> .78), VIF’s were well be-
low 4, and condition indices were lower than 16. Based on this evidence, it was deter-
mined that multicollinearity would pose no threat when estimating the citizen-level
regression models (Fox, 1991).
4. Although the observed correlations suggest possible issues of multicollinearity,
relying exclusively on bivariate associations presents limitations (Belsley et al., 1980, pp.
92-93; Berry & Feldman, 1985, p. 45). Therefore, mean perceived incivility was regres-
sed on the council district variables to explore whether multicollinearity existed. Similar to
the citizen-level diagnostic results, the findings revealed that the tolerance statistics were
sufficiently high (> .72), VIF’s were well below 4, and condition indices were lower than 4.
5. Reliability (+) is defined as ,[*00/(*00 !'
2/nk)]/K, the average of council district
specific reliabilities across the set of aggregates (K= 10). Reliability is a function of the
sample size (nk) in each council district and the proportion of the total variance between
council districts (*00) relative to the amount within council districts ('2).
REFERENCES
Anderson, E. (1990). Streetwise. Chicago, IL: University of Chicago Press.
Aneshenal, C. S. & Sucoff, C. A. (1996). The neighborhood context of adolescent men-
tal health. Journal of Health and Social Behavior, 37, 293-310.
Banji, M. R. (2002). Social psychology of stereotypes. In N. J. Smelser & P. B. Baltes
(Eds.), International encyclopedia of the social and behavioral sciences (pp. 15100-
15104). Oxford: Elsevier.
22 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Bank, L., Dishion, T., Skinner, M., & Patterson, G. R. (1989). Method variance in
structural equation modeling: Living with “glop.” In G. R. Patterson (Ed.), Aggres-
sion and depression in family interactions (pp. 25-40). Hillsdale, NJ: Lawrence
Erlbaum.
Barron, D. N. (1992). The analysis of count data: Overdispersion and autocorrelation.
Sociological Methodology, 22, 179-220.
Bellair, P. E. (2000). Social interaction and community crime. Examining the impor-
tance of neighbor networks. Criminology, 35, 677-701.
Belsley, D. A., Kuh, E., & Welsch, R. E. (1980). Regression diagnostics: Identifying
influential data and sources of collinearity. New York, NY: John Wiley.
Berry, W. D. & Feldman, S. (1985). Multiple regression in practice.NewburyPark,CA:
Sage Publications.
Blakely, T. A. & Woodard, A. J. (2000). Ecological effects in multi-level studies. Jour-
nal of Epidemiology and Community Health, 54, 367-374.
Bryk, A. S. & Raudenbush, S. W. (1992). Hierarchical linear models: Applications
and data analysis methods. Newbury Park, CA: Sage Publications.
Burton, C. E. (2004). Segregation and Latino homicide victimization. American Jour-
nal of Criminal Justice, 29, 21-36.
Chambliss, W. J. (1994). Policing the ghetto underclass: The politics of law and law
enforcement. Social Problems, 41, 177-194.
Charles, C. Z. (2003). The dynamics of racial residential segregation. Annual Review of
Sociology, 29, 167-207.
Chiricos, T., McEntire, R., & Gertz, M. (2001). Perceived racial and ethnic com-
position of neighborhood and perceived risk of crime. Social Problems,48,
322-340.
Clark, W. W. (1992). Residential preferences and residential choices in a multiethnic
context. Demography, 29, 451-466.
Duncan, G. J. & Aber, J. L. (1997). Neighborhood models and measures. In
J. Brooks-Gunn, G. J. Duncan, & J. L. Aber (Eds.), Neighborhood poverty: Con-
text and consequences for children (pp. 62-78). New York, NY: Russell Sage
Foundation.
Ferraro, K. F. (1994). Fear of crime: Interpreting victimization risk. Albany, State Uni-
versity of New York Press.
Ferraro, K. F. & LaGrange, R. I. (1987). The measurement of fear of crime. Sociologi-
cal Inquiry, 57, 70-101.
Garofalo, J. & Laub, J. (1978). The fear of crime: Broadening our perspective. Victim-
ology, 3, 242-253.
Guzman, B. (2001). Census 2000 brief: The Hispanic population. Washington, DC:
U.S. Department of Commerce Economics and Statistics Administration.
Hagan, J. & Peterson, R. D. (1995). Criminal inequality in America: Patterns and con-
sequences. In J. Hagan & R. D. Peterson (Eds.), Crime and inequality (pp. 14-36).
Stanford, CA: Stanford University Press.
Holmes, M. D. (2003). Ethnicity, concentrated disadvantage, and perceived risk of vic-
timization. Journal of Ethnicity in Criminal Justice, 1, 1-20.
Hunter, A. (1978). Symbols of incivility. Paper presented at the Annual Meeting of
the American Society of Criminology. Dallas, TX.
Cancino et al. 23
Jacobs, J. (1961). The death and life of great American cities. New York, NY: Random
House.
Kreft, I. & De Leeuw, J. (1998). Introducing multilevel modeling. Thousand Oaks, CA:
Sage Publications.
Land, K. C., McCall, P. L., & Cohen, L. E. (1990). Structural covariates of homicide
rates: Are there any invariances across time and social space? American Journal of
Sociology, 92, 922-963.
Lane, J. & Meeker, J. W. (2003). Fear of gang crime: A look at three theoretical mod-
els. Law & Society Review, 37, 425-456.
______. (2004). Social disorganization perceptions, fear of gang crime, and behavioral
precautions among Whites, Latinos, and Vietnamese. Journal of Criminal Justice,
32, 49-62.
______. (2005). Theories and fear if gang crime among Whites and Latinos: A replica-
tion and extension of prior research. Journal of Criminal Justice, 33, 627-641.
Lee, M. R. & Ousey, G. C. (2005). Institutional access, residential segregation, and ur-
ban black homicide. Sociological Inquiry, 75, 31-54.
Lee, S. K. & Ulmer, J. T. (2000). Fear of crime among Korean Americans in Chicago
communities. Criminology, 38, 1173-1206.
Lewis, D. A. & Maxfield, M. G. (1980). Fear in the neighborhoods: An investigation
of the impact of crime. Journal of Research in Crime and Delinquency,17,
160-189.
Lieberson, S. (1981). An asymmetrical approach to segregation. In V. Robinson & S.
Smith (Eds.), Ethnic segregation in cities (pp. 61-82). London: Croom Helm.
Long, S. (1997). Regression models for categorical and limited dependent variables.
Thousand Oaks, CA: Sage.
Loury, G. (2002). The anatomy of racial inequality.Cambridge,MA:HarvardUniver-
sity Press.
Markowitz, F. E., Bellair, P. E., Liska, A. E., & Liu, J. (2001). Extending social disor-
ganization theory: Modeling the relationships between, cohesion, disorder, and
fear. Criminology, 39, 293-319.
Marotta, S. A. & Garcia, J. G. (2003). Latinos in the United States. Hispanic Journal of
Behavioral Sciences, 25, 25-14.
Massey, D. S. & Denton, N. A. (1988). Suburbanization and segregation in U.S. metro-
politan areas. American Journal of Sociology, 96, 329-357.
______. (1993). American apartheid: Segregation and the making of the underclass.
Cambridge, MA: Harvard University Press.
Mayhew, H. (1862). London labor and the London poor. London: Griffin, Bohn.
Mok, M. & Flynn, M. (1998). Effect of Catholic school culture of students’ achieve-
ment in higher school certificate examinations: A multilevel path analysis. Educa-
tional Psychological, 18, 409-432.
Mosher, C. (2001). Predicting drug arrest rates: Conflict and social disorganization
perspective. Crime and Delinquency, 47, 84-104.
Peterson, R. D. & Krivo, L. J. (1993). Racial segregation and black urban homicide.
Social Forces, 71, 1001-1026.
Quillian, L. & Pager, D. (2001). Black neighbors, high crime? The role of racial stereo-
types in evaluations of neighborhood crime. American Journal of Sociology, 107,
717-767.
24 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
Reisig, M. D. & Cancino, J. M. (2004). Incivilities in nonmetropolitan communities:
The effects of structural constraints, social conditions, and crime. Journal of Crimi-
nal Justice, 32, 15-29.
Reisig, M. D. & Parks, R. B. (2004). Can community policing help the truly disadvan-
tage? Crime & Delinquency, 50, 139-167.
Reiss, A. J., Jr. (1973). Monitoring the quality of criminal justice systems. In A.
Campbell & P. Converse (Eds.), The human meaning of social change (pp. 388-
403). New York, NY: Russell Sage Foundations.
Rountree, P. W., Land, K. C., & Miethe, T. D. (1994). Macro-micro integration in the
study of victimization: A hierarchical logistic model analysis across Seattle neigh-
borhoods. Criminology, 32, 387-405.
Sampson, R. J. & Groves, W. B. (1989). Community structure and crime: Testing so-
cial disorganization theory. American Journal of Sociology, 94, 774-802.
Sampson, R. J. & Jeglum-Bartusch, D. (1998). Legal cynicism and (subcluture?) toler-
ance of deviance: The neighborhood context of racial differences. Law and Society
Review, 32, 777-804.
Sampson, R. J. & Raudenbush, S. W. (1999). Systemic social observation of public
spaces: A new look as disorder in urban neighborhoods. American Journal of Soci-
ology, 105, 603-651.
______. (2004). Seeing disorder: Neighborhood stigma and the social construction of
“broken windows.” Social Psychology Quarterly, 67, 319-342.
Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhood and violent
crime: A Multilevel study of collective efficacy. Science, 277, 918-924.
Shaw, C. R. & McKay H. D. (1942). Juvenile delinquency and urban areas: A study of
rates of delinquency in relation to differential characteristics of local communities
in American cities. Chicago, IL: University of Chicago Press.
Shihadeh, E. S. & Flynn, N. (1996). Segregation and crime: The effect of black social
isolation on the rates of black urban violence. Social Forces, 74, 1325-1352.
Skogan, W. (1990). Disorder and decline: Crime and the spiral of decay in American
cities. New York, NY: Free Press.
Skogan, W. G. & Maxfield, M. G. (1981). Coping with crime: Individual and neigh-
borhood reactions. Beverly Hills, CA: Sage Publications.
Skogan, W. G. & Steiner, L. (2004). Crime, disorder, and decay in Chicago’s Latino
community. Journal of Ethnicity in Criminal Justice, 2, 7-26.
Small, M. L. & Newman, K. (2001). Urban poverty after the truly disadvantage: The
rediscovery of the family, the neighborhood, and culture. Annual Review of Sociol-
ogy, 27, 23-45.
Stucky, T. D. (2005). Local politics and police strength. Justice Quarterly,22,
139-169.
Taylor, R. B. (2001). Breaking away from broken windows. Boulder, CO: Westview.
Taylor, R. B. & Covington, J. (1993). Community structural change and fear of crime.
Social Problems, 40, 375-395.
Taylor, R. B. & Shumaker, S. (1990). Local crime as a natural hazard: Implications for
understanding the relationship between disorder and fear. American Journal of
Community Psychology, 18, 619-641.
Cancino et al. 25
Warr, M. (2000). Fear of crime in the United States: Avenues for research and policy.
In D. Duffee (Ed.), Measurement and analysis of crime and justice (pp. 452-489).
Washington, DC: National Institute of Justice.
Welsh, W. N., Greene, J. R., & Jenkins, P. H. (1999). School disorder: The influence of
individual, institutional, and community factors. Criminology, 37, 73-115.
Wilson, W. J. (1987). The truly disadvantaged: The inner city, the underclass, and pub-
lic policy. Chicago, IL: The University of Chicago Press.
Wilson, J. Q. & Kelling, G. (1982). Broken windows. Atlantic Monthly, 211, 29-38.
Received: June 2006
Accepted: August 2006
doi:10.1300/J222v05n01_01
26 JOURNAL OF ETHNICITY IN CRIMINAL JUSTICE
