Transcript
Page 1: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

at SciVerse ScienceDirect

Journal of Environmental Psychology 32 (2012) 43e49

Contents lists available

Journal of Environmental Psychology

journal homepage: www.elsevier .com/locate/ jep

Making neighborhoods safer: Examining predictors of residents’ concerns aboutneighborhood safety

Ronald O. Pitner a,*, ManSoo Yu b, Edna Brown c

aCollege of Social Work, University of South Carolina, USAb School of Social Work and Public Health Program, University of Missouri, USAcDepartment of Human Development and Family Studies, University of Connecticut, USA

a r t i c l e i n f o

Article history:Available online 1 October 2011

Keywords:Neighborhood violencePerceptionsIncivilityCollective efficacyCrimeBroken windows

* Corresponding author. Tel.: þ1 803 777 6797.E-mail addresses: [email protected], [email protected]

0272-4944/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.jenvp.2011.09.003

a b s t r a c t

This study examined what factors best predict residents’ concerns about neighborhood safety. One-hundred and twenty-two participants were selected from a large, Midwestern metropolitan area. Allparticipants lived in high crime areas. Participants completed a 22-item questionnaire that assessed theirperceptions of neighborhood safety and vigilance. These items were clustered as: 1) Community care andvigilance, 2) neighborhood safety concerns, 3) physical incivilities, and 4) social incivilities. Police crimedata were also used in the analyses. Our findings suggest that aspects of the broken window theory,collective efficacy, and place attachments/territoriality play a role in affecting residents’ concerns aboutneighborhood safety.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Community violence continues to plague the lives of residentswho live in our nation’s most vulnerable, high-crime neighbor-hoods. Such violence leads to resident disengagement, and oftenerodes away their overall perceptions of neighborhood safety (e.g.,Taylor, 2002). Understanding the dynamics that affect residents’perceptions of neighborhood safety in these high crime areas couldbe one key to helping residents work together to deter crime. Itcould also lead to practical implications for improving neighbor-hood conditions for our most vulnerable communities. The purposeof this study was to examine the predictors of concerns aboutneighborhood safety.

Which theoretical frameworks are most appropriate to examineresidents’ feelings of safety in these vulnerable, high crime neigh-borhoods? Prior research has been mixed on this question. Whilesome studies have concluded that broken window theory isappropriate to explain predictors of concerns about neighborhoodsafety (Bechtel & Churchman, 2002; Colquhoun, 2004; Day, 1994;Goldstein, 1994; Greenberg, Rohe, & Williams, 1982; Newman &Franck, 1980; Perkins, Meeks, & Taylor, 1992; Roman & Chalfin,2008; Skogan, 1976; Taylor, 1997; White, 1990), other studies indi-cate that collective efficacy (e.g., Sampson, 2004; Sampson & Graif,2009; Sampson & Raudenbush, 1999; Shamard, 2009) and/or

om (R.O. Pitner).

All rights reserved.

place attachments/territoriality frameworks (e.g., Brown, Perkins, &Brown, 2003; 2004) are more appropriate.

1.1. Toward the broken window theory hypothesis

Physical and social cues serve as implicit markers for unsafe andviolence-prone neighborhoods (Astor, Meyer, & Pitner, 2001;Brantingham & Brantingham, 1981; Eck & Weisburd, 1995;Newman & Franck, 1980; Perkins, Florin, Rich, Wandersman, &Chavis, 1990; Perkins et al., 1992; Pitner & Astor, 2008; Pitner, Yu,& Brown, 2011; Skogan, 1976; Taylor, 1994, 1997; Taylor &Gottfredson, 1986; White, 1990). These areas are characterized byphysical incivilities (e.g., vandalism, graffiti, and debris in yards)and social incivilities (e.g., noisy neighbors, prostitution, drugtrafficking, and gang-related activity) (Perkins, Wandersman, Rich,& Taylor, 1993). Research suggests that increased neighborhoodincivilities invoke perceptions of crime and disorder among resi-dents and potential offenders, which could potentially lead tohigher neighborhood crime (Perkins et al., 1992; Taylor, 1999;Taylor & Gottfredson, 1986; Taylor, Gottfredson, & Brower, 1984).

Areas that have high rates of social and physical incivilities maybe thought of as undefined public spaces (Cisnernos, 1995;Newman, 1973, 1995; Newman & Franck, 1980). The term unde-fined refers to areas that lack ownership. In such areas, residentsfeel that it is not their responsibility to monitor or maintain them.As a consequence of this lack of ownership, the rates of incivilitiesin these areas increase, leading to heightened perceptions oflawlessness and crime. Politicians and police departments often

Page 2: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e4944

refer to this concept as the “broken window” theory. This theoryposits that a sequence of events typically takes place in undefinedpublic spaces:

Evidence of decay (accumulated trash, broken windows, dete-riorating building exteriors) remains in the neighborhood fora reasonably long period of time. People who live and work inthe area feel more vulnerable and begin to withdraw. Theybecome less willing to intervene to maintain public order or toaddress physical signs of deterioration. Sensing this.offendersbecome bolder and intensify their harassment and vandalism.Residents become yet more fearful and withdraw further fromcommunity involvement and upkeep. This atmosphere thenattracts offenders from outside the area, who sense that it hasbecome a vulnerable and less risky site for crime (Wilson &Kelling, 1982).

Several studies have documented the relation between unde-fined spaces and residents’ perceptions of danger (Day, 1994;Goldstein, 1994; Greenberg et al., 1982; Newman & Franck, 1980;Perkins et al., 1992; Pitner & Astor, 2008). For example, Pitner andAstor (2008) showed that children who live in high crime areaswere more likely to make harm attributions about places that theyperceived as undefined. However, a question remains aboutwhether it is the perceived presence of social incivilities, theperceived presence of physical incivilities, or the perceived pres-ence of both that have the greatest associationwith concerns aboutneighborhood safety.

1.2. Toward the collective efficacy hypotheses

Sampson and Raudenbush (1999) contend that althoughresearch shows the connection between neighborhood incivilitiesand fear of crime, this relationship is a spurious one. For them,collective efficacy underlies both neighborhood incivilities andperceptions of neighborhood crime. Collective efficacy is anagency-oriented perspective focused on residents taking an activerole in shaping their neighborhood communities (Sampson &Raudenbush, 1999). This approach, by default, requires a highlevel of community cohesion. Such cohesion militates againstresidents’ fear of neighborhood crime (Sampson, 2004; Sampson &Raudenbush, 1999). More specifically, decreases in collective effi-cacy increases perceptions of incivilities, and increases perceptionsof neighborhood crime. Voluminous studies have used the collec-tive efficacy framework, and suggest an inverse relationshipbetween levels of community cohesion and perceptions of neigh-borhood safety (e.g., Ferguson & Mindel, 2007; Franzini, Caughy,Spears, & Esquer, 2005; Sampson, 2004; Sampson & Graif, 2009;Sampson, Morenoff, & Earls, 1999; Sampson & Raudenbush, 1999;Shamard, 2009; Wells, Schafer, Varano, & Bynum, 2006).

1.3. Toward the place attachment/territoriality hypotheses

Place attachment is a term used in environmental psychologythat focuses on the bonds between residents and their social andphysical settings (Brown, Perkins, & Brown, 2004). Related to this isthe term territoriality, which is defined as informal social controlthat consists of residents displaying pride and ownership of theirproperty through physical markers (e.g., gardens, planting shrubsor flowers) (Perkins et al., 1993). In this study, we used the termsplace attachment and territoriality together, as research shows thatresidents who display high levels of place attachment often showstrong territorial commitment (e.g., Harris & Brown, 1996). In manyways, place attachment and territoriality are similar concepts tocollective efficacy in that they all emphasize the important role thatneighborhood cohesion plays in residents’ perceptions of safety. In

fact, higher levels of place attachment/territoriality are associatedwith higher levels of collective efficacy (e.g., Long & Perkins, 2007;Tester, Ruel, Anderson, & Oakley, 2011). However, place attach-ment/territoriality differ from collective efficacy in that they alsoplace emphasis on residents’ feelings of pride in their home andneighborhood appearance, and on residents’ tenure in theirneighborhoods (Brown &Werner, 1985; Brown, Perkins, & Douglas,1992). Brown et al. (2004) suggest that decreased place attach-ments/territoriality can directly increase concerns about neigh-borhood safety.

1.4. Present study

The current study used the concept of “community care andvigilance,” which we conceptualized as a proxy for both collectiveefficacy and place attachment/territoriality. Borrowing both fromaspects of collective efficacy (i.e., willingness to take a stance against“normviolating” behaviors) and place attachment/territoriality (i.e.,neighborhood pride), we defined community care and vigilance asresidents’ pride in their neighborhoods and sense of community, aswell as their willingness to take action to protect their neighbor-hoods. Consistent with previous research (e.g., Sampson, 2004;Sampson & Graif, 2009; Wells et al., 2006), we hypothesize thatheightened community care and vigilance will be inversely associ-ated with residents’ concerns about neighborhood safety.

Previous research has contributed significantly to the knowl-edge base on perceptions of neighborhood safety. Nevertheless,there is a paucity of research that comprehensively examines therelationship that community care and vigilance and perceivedneighborhood incivilities have on neighborhood safety concerns forvulnerable residents in high-crime neighborhoods. Specifically, is itthe sole presence of perceived physical incivilities, the sole pres-ence of perceived social incivilities, or a combined presence ofincivilities that have the greater association to safety concerns? Weexamined these issues in this study. In addition, because actualneighborhood crime (i.e., crimes against property/person) can havean important effect on residents’ perceptions of neighborhoodsafety, we examined whether crime played a stronger role thancommunity care and vigilance and perceived incivilities in pre-dicting residents’ concerns about neighborhood safety.

1.5. Hypotheses

The purpose of this study is to lookmore comprehensively at thepredictors of concerns about neighborhood safety among vulner-able populations. Following directly from the aforementionedtheories and supporting research, four hypotheses were tested inthis study. First, using a broken window theory paradigm, wehypothesized that both perceived physical and social incivilitieswould predict residents’ concerns about neighborhood safety. Wewere uncertain whether this would vary by the type of incivility(i.e., physical vs. social), or whether it would be an effect of thecombined presence of any incivility. Thus, no specific hypotheseswere stated regarding this. Second, using a collective efficacy andplace attachment/territoriality paradigm, we hypothesized thatcommunity care and vigilance would be a predictor of residents’concerns about neighborhood safety. Third, we hypothesized thatactual crime rate (i.e., police crime reports) would predict safetyconcerns, and that this would vary by type of crime (i.e., crimesagainst property vs. crimes against person). Finally, given thatperceived incivilities can erode neighborhood social cohesion andcollective efficacy (e.g., Wells et al., 2006), we hypothesized thatincivilities (both physical and social) would be the strongestpredictor of concerns about neighborhood safety among vulnerableresidents.

Page 3: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e49 45

2. Methods

2.1. Sample and participant selection

This study received university-level approval from the InternalReview Board. The study was conducted in an urban city withina largeMidwestern, metropolitan area. Therewere 122 participantsin this study, among whom the majority were female (82.8%).Seventy-six (65.0%) participants were African-American and 41(35.0%) were White. On average, participants lived in their neigh-borhoods for 16.6 years (SD ¼ 15.1), with a range of 0.25e52 years.Moreover, the average age was 54.1 (SD ¼ 17.7), ranging from 20 to89. Annual household income ranged from 1 (less than $10,000) to7 (more than $75,000), with the mode lying under $10,000 (43%).Detailed characteristics of the study sample are provided in Table 1.

Participants were selected because they received free servicesfrom a non-profit organization that focused on helping residentswho live in high crime areas feel safer in their neighborhoods. Inparticular, residents were assisted with the following homeprotection devices: dead-bolt lock(s), window-ventilators, doorpeephole(s), and basement window bars. We examined a 6-monthperiod (April 2002 to September 2002) for which the non-profitorganization provided services to residents. During that period,496 residents received services.

The target city was segmented into 9 police districts. Given this,we employed a stratified random sampling procedure. Within eachof these 9 districts, the sample was stratified by month of servicedelivery (April through September). Residents were then selectedrandomly by strata. A total of 24 residents were selected randomlyper district, with the exception of District 4 and District 9. For thesetwo districts, few residents received services from the non-profitorganization (2 in each district). Thus, all 4 residents wereselected for the sample. One-hundred and seventy-two residentswere selected, and at least 60.0% of the selected sample partici-pated from each police district. Overall, 122 residents participatedin the study, which was 70.9% of the selected sample.

Table 1Descriptive information.

Overall sample(N ¼ 122)

Male (n ¼ 21)

M SD % M SD %

DemographicsAge 54.1 17.7 52.1 17.8RaceWhite 35.0% 42.9%African Americans 65.0% 57.1%

Years of living in current resident 16.6 15.1 17.5 15.9Annual income 2.4 1.7 2.7 2.2

$10,000 or less (1) 43.3% 42.1%$10,001e$20,000 (2) 22.5% 26.3%$20,001e$30,000 (3) 10.8% 5.3%$30,001 or more (4,5,6,7) 23.3% 26.3%

Perceived environmental crime variablesCommunity care and vigilance 22.0 5.7 24.2 5.9Incivility 4.7 2.6 4.2 2.8Physical incivility 2.9 1.9 2.4 1.8Social incivility 1.8 1.2 1.8 1.3

Crime variablesCrimes against person 3.0 1.3 2.4 1.2Crimes against property 1.7 0.6 1.7 0.5

Neighborhood safety concerns 6.1 2.4 4.8 2.2

a Because of the multiple comparisons conducted in this study, the adjusted criticalindividual number of tests conducted in this study (k¼ 10). Thus, the null hypothesis woustudy variables between male and female residents were significant.

b The Bonferroni correction was not applied to the 2 by 2 Chi-square test because it h

2.2. Interview procedures

We used a mixed data collection procedure, which consisted ofmailed surveys and telephone interviews. Residents were initiallymailed surveys to be completed and returned. However, if they didnot return the survey within 3 weeks, a trained researcher con-tacted them by telephone and they then completed the survey onthe telephone. Given that we used a stratified random samplingprocedure, our study participants were representative of theoverall group of residents who received services during the monthsof April and September.

We also collected police crime data for the 9 districts. The crimecategories (i.e., total crimes against persons and total crimes againstproperty) were coded, where lower numbers denote lower levels ofcrime and higher numbers denote higher levels.

2.3. Instrument

The questionnaire was created conjointly by members of ourresearch team and members from the non-profit organization.Participants completed this 22-item questionnaire that consisted ofquestions pertaining to neighborhood safety concerns (2 items),perceived environmental crime variables: community care andvigilance (6 items), physical incivilities (6 items) and social inci-vilities (4 items), as well as demographic information. Safetyconcerns assessed the degree to which the person felt unsafe in hisor her neighborhood, and the degree he or she felt it was unsafe forchildren and adults. Community care and vigilance items were:people watch out for each other in my neighborhood; if I witnesseda crime in my neighborhood, I would report it; people in myneighborhood care about the area; my neighborhood feels likea community; I feel good about living in my neighborhood; andthere is not a lot of crime in my neighborhood. Physical incivilityitems pertained to decayed and run-down buildings, abandonedbuildings, vacant lots, debris in the streets, graffiti, and poorlighting. Finally, social incivility items pertained to gangs, drug

Female (n ¼ 101) Analysis

M SD % t/c2 df Unadjusted p p withBonferroniadjustmena

Range

54.6 17.9 0.57 120 0.56 1.00 20e890.69 1 0.41 N.A.b 0e1

33.3%66.7%

16.2 15.0 0.34 119 0.74 1.00 0.25e522.3 1.6 1.03 118 0.30 1.00 1e7

43.5%21.8%11.9%22.8%

21.5 5.6 2.13 120 0.03 0.35 5e304.8 2.6 0.93 120 0.36 1.00 0e103.0 1.9 1.17 120 0.24 1.00 0e61.9 1.2 0.18 120 0.86 1.00 0e4

3.1 1.3 2.29 109 0.02 0.24 1e51.7 0.7 0.05 109 0.96 1.00 1e56.3 2.4 2.65 119 0.009 0.09 2e10

value for rejection of the null hypothesis was calculated as 0.5/k, where was theld have been rejected if p< .005. With use of this criterion, none of differences in the

as only one degree of freedom.

Page 4: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

Table 2Percentage of incivilities (N ¼ 122).

Overall sample Male Female

Any incivility 96.7% 95.2% 97.0%Any physical incivility 86.1% 85.7% 86.1%Decayed and run-down buildings 59.0% 47.6% 61.4%Abandon buildings 58.2% 47.6% 60.4%Poor lighting 54.1% 52.4% 54.5%Vacant lots 50.0% 38.1% 52.5%Debris 41.0% 28.6% 43.6%Graffiti 24.6% 28.6% 23.8%

Any social incivility 84.4% 81.0% 85.2%Drug traffickers and addicts 71.3% 66.7% 72.3%Gangs 50.8% 52.4% 50.5%Nuisance and problem neighbors 39.3% 42.9% 38.6%Homeless people 23.8% 19.1% 24.8%

Table 3Correlations of study variables (N ¼ 122).

(A) (B) (C) (D) (E) (F) (G)

(A) Age 1.00(B) Years of living in

current resident0.57*** 1.00

(C) Annual income �0.30*** �0.11 1.00(D) Community care

and vigilance0.16 0.06 �0.14 1.00

(E) Physical incivility �0.12 �0.02 �0.17 �0.24** 1.00(F) Social incivility �0.14 �0.14 0.11 �0.24** 0.42*** 1.00(G) Neighborhood

safety concerns�0.03 0.06 �0.08 �0.33*** 0.34*** 0.26** 1.00

Note: Only continuous variables were included for correlation analysis.**p < .01, ***p < .001.

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e4946

traffickers, homeless people, and nuisance and problem neighborsmaking the neighborhood unsafe. Itemswere rated on a Likert scaleranging from “strongly disagree ¼ 1” to “strongly agree ¼ 5.” Thedata demonstrated acceptable internal consistency reliability foreach of the summed measures: neighborhood safety concerns(alpha ¼ 0.74), physical incivilities (alpha ¼ 0.72), social incivilities(alpha ¼ 0.52), and community care and vigilance (alpha ¼ 0.79) inthis sample.

2.4. Data analysis

Chi-square and independent t-tests were performed to examineif demographic variables (i.e., race, age, annual income, and lengthof residency), perceived environmental crime variables, actualcrime variables, and concerns about neighborhood safety differedby gender. Univariate analyses were performed to estimate thefrequency of physical and social incivilities. Zero-order correlationswere employed to assess associations among all study variables.Finally, multiple linear regression analyses were conducted toinvestigate which of the independent variables (i.e., demographics,perceived environmental crime variables, and crime variables)significantly predicted the dependent variable (i.e., neighborhoodsafety concerns). No multicollinearity problems were observed inthat all tolerance values exceeded 0.25 (Fox, 1991). Analyses wereconducted using SAS 9.2 (SAS Institute Inc., 2009). In addition, weused statistical power to determine our sample size (Cohen, 1988).Based on an anticipated large effect size (f2 ¼ 0.30), alpha ¼ 0.05and number of predictors at 8, minimum total sample size wascalculated at 84 participants with a statistical power of 0.95 (Faul,Erdfelder, Lang, & Buchner, 2007). Thus our sample size issatisfactory.

3. Results

3.1. Descriptive information

Descriptive information of participants is presented in Table 1. Inthis study, perceived environmental crime variables consisted ofitems assessing community care and vigilance, and physical andsocial incivilities. The higher an individual scored on these items,the stronger the endorsement of the item. Overall, participantsreported a mean score of 22.0 (SD ¼ 5.7) for community care andvigilance; scores ranged from 5 to 30. Participants also indicated anaverage score of 4.7 (SD ¼ 2.6) on items assessing incivilities, witha range of 0e10. The average number of physical incivilitiesendorsed was 2.9 (SD ¼ 1.9), with a range of 0e6; and the averagenumber of social incivilities endorsed was 1.8 (SD ¼ 1.2), witha range of 0e4.

In terms of crime variables, both crimes against person andcrimes against property ranged from 1 to 5, where higher numbersindicated higher rates of crime. The average level for crimes againstpersonwas 3.0 (SD¼ 1.3), while the average level for crimes againstproperty was 1.7 (SD ¼ 0.6). The mode for both crime variables waslevel 2 (crimes against person at 50% and crimes against property at62%). Lastly, safety concerns consisted of items assessing residents’perceived concerns about neighborhood safety; higher scoresrepresented greater endorsement of the item (i.e., greater feelingsof being unsafe). Overall, participants reported a mean score of 6.1(SD ¼ 2.4) for neighborhood safety concerns, with a range of 2e10.

Table 2 presents the percentage of participants who identifiedspecific incivilities. An overwhelming percentage of participants(96.7%) identified at least one type of incivility that they perceivedas making their neighborhood less safe. Regarding specific incivil-ities, 86.1% identified at least one physical incivility, and 84.4%identified at least one social incivility that made their

neighborhood unsafe. Among the subtypes of physical incivilities,a greater percentage of participants reported that decayed build-ings made their neighborhoods less safe (59%), followed byabandon buildings (58%), poor lighting (54%), vacant lots (50%),debris (41%), and graffiti (25%). Among the subtypes of social inci-vilities, a greater percentage reported that drug traffickers andaddicts made their neighborhoods less safe (71%), followed bygangs (51%), nuisance and problem neighbors (39%), and homelesspeople (24%).

3.2. Bivariate correlations

Table 3 presents correlations of the seven continuous variablesin the study. Higher levels of concerns about neighborhood safetywere associated positively with levels of perceived physical andsocial incivilities, and negatively with community care and vigi-lance. Higher levels of community care and vigilance were associ-ated with lower levels of social and physical incivilities. There wasa positive correlation between social and physical incivilities. Agewas correlated positively with years of living in current residentand negatively with annual income. These correlations werestatistically significant, with r values ranging from 0.24 to 0.57.

3.3. Multiple linear analyses for predicting neighborhood safetyconcerns

We ran 4 different multiple regression models to identify whichindependent variables (i.e., demographics, perceived environ-mental crime variables, and crime variables) best predictedconcerns about neighborhood safety (Table 4). Model 1 [F (6,112) ¼ 5.23; p < .001] contained demographic and perceivedenvironmental crime variables, and explained 22% of the variancein concerns about neighborhood safety. As can be seen, lower

Page 5: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

Table 4Linear regression analyses for predicting neighborhood safety concerns.

Predictor Model 1(N ¼ 119) Model 2(N ¼ 109) Model 3 (N ¼ 109) Model 4 (N ¼ 109)

b t b t b t b t

DemographicsFemale 0.16 1.92y 0.20 2.18* 1.16 1.81y 1.16 1.82Age 0.03 0.36 �0.02 �0.15 0.06 0.63 0.06 0.63Annual income �0.07 �0.86 0.01 0.08 �0.04 �0.42 �0.04 �0.43

Perceived environmental crime variablesCommunity care and vigilance �0.24 �2.64** �0.21 �2.17* �0.21 �2.21*Any incivility 0.25 2.75**Physical incivility 0.21 2.18* 0.19 1.85ySocial incivility 0.12 1.20 0.12 1.14

Crime variablesCrimes against person 0.11 1.09 0.02 0.21 0.02 0.22Crimes against property 0.23 2.31* 0.22 2.33* 0.22 2.36*

Model F 5.23*** 3.15* 4.25*** 4.91***R-Square 0.22 0.13 0.25 0.25

Note. b ¼ standardized estimate.yp < .10, *p < .05, **p < .01, ***p < .001.

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e49 47

community care and vigilance and higher perceived physical inci-vilities predicted higher concerns about neighborhood safety.Model 2 [F (5, 103) ¼ 3.15; p ¼ 010], which contained only demo-graphic and crime variables, revealed that crimes against propertypositively predicted concerns about neighborhood safety. Model 2explained 13% of the variance.

Model 3 [F (8, 100) ¼ 4.25; p < .001] included all independentvariables that were contained in Models 1 and 2. More specifically,the independent variables were demographic variables, perceivedcrime variables, and crime variables. Model 3 explained 25% of thevariance in concerns about neighborhood safety. Once all variableswere entered into the model, community care and vigilance andcrimes against property remained significant predictors ofconcerns about neighborhood safety. Physical incivilities margin-ally predicted such concerns (b ¼ 0.19, p ¼ .07).

Given that we hypothesized that perceived incivilities would bethe strongest predictor of concerns about neighborhood safety, weran the last model, which included the following independentvariables: demographic variables, crime variables, and “any inci-vility” (which denoted that participants indicated that theirneighborhoods are unsafe because of the presence of at least onephysical or social incivility). As shown in Model 4 [F (7, 101) ¼ 4.91;p< .001], higher reports of any incivility (p¼ .007), as well as lowercommunity care and vigilance, and higher crimes against propertysignificantly predicted concerns about neighborhood safety. Model4 explained 25% of the variance.

4. Discussion

The purpose of this study was to examine what variables bestpredict residents’ concerns about neighborhood safety. Our find-ings support the second and third hypotheses, and partially supportour first and fourth hypotheses. Specifically, we hypothesized thatcommunity care and vigilance (hypothesis 2) and crime (hypothesis3) would predict neighborhood safety concerns, which was thecase. We also hypothesized that both physical and social incivilitieswould predict concerns about neighborhood safety in hypothesis 1.This hypothesis was partially supported because only physicalincivility predicted safety concerns. Findings from the multiplelinear regression analysis (Model 4) partially supported hypothesis4, indicating that “any” incivility (i.e., the presence of at least onephysical or social incivility) was the strongest predictor concernsabout neighborhood safety.

Overall, our findings are consistent with previous research inshowing that higher levels of concern about neighborhood safety

are correlated with higher levels of perceived physical incivilities(e.g., Astor et al., 2001; Colquhoun, 2004; Perkins et al., 1992; Pitner& Astor, 2008; Taylor, 1997), higher crimes against person andproperty (e.g., Taylor, 2002), and lower levels of community care andvigilance (e.g., Pitner et al., 2011; Sampson & Graif, 2009; Shamard,2009). Unlike these previous studies, we uniquely examined theindependent associations that demographics, community care andvigilance, incivilities, and crime have on concerns about neighbor-hood safety in a sample of residents living in high crimeneighborhoods.

Specifically, two perceived environmental crime variables weresignificant predictors of concerns about neighborhood safety:Community care and vigilance and perceived incivilities. Consistentwith research on collective efficacy and place attachments/terri-toriality, our findings suggest that when residents express prideand a sense of community in their neighborhoods, as well asa willingness to take action to protect their neighborhood, they areless likely to have concerns about neighborhood safety. This couldsuggest that residents’ feelings and overall perceptions of collectiveengagement may militate against concerns about neighborhoodsafety. Future research should examine this issue more compre-hensively. Was there a relationship between the presence of anyincivility and levels of residents’ concerns about neighborhoodsafety? Overall, most participants (97%) in this study perceived thepresence of any incivility as a potential issue that makes theirneighborhoods less safe. In fact, when this variable was enteredinto the regression model, it was the strongest predictor of resi-dents’ concerns about neighborhood safety.

Our findings suggest that heightened levels of perceived phys-ical incivilities increase residents’ concerns about neighborhoodsafety. It is interesting to note that social incivilities did not predictconcerns about neighborhood safety in this study. One compellingreason why this may be the case is that social incivilities are notconstant fixtures in neighborhoods. Specifically, sometimes drugtraffickers, homeless people, and/or gangs may be present, othertimes they may not, whereas the presence of neighborhood phys-ical incivilities (e.g., an abandon building) is often much moreconstant. Thus, physical incivilities may provide a better indicatorof the environmental conditions of the neighborhood than socialincivilities. A second reason may have to do with the age of theresidents in this study. Given that gang members are often youngerpeople from the neighborhood, and often less known to olderadults, it could be that older adults are not aware of “gang” or drug-related activities in which these younger residents engage. A finalreason could be that many activities related to social incivilities are

Page 6: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e4948

more common at night. Yet, older adult residents may be morelikely to venture outside during daylight hours, making them lessaware of their neighborhood social incivilities.

In our study, community care and vigilance had a strongerassociation than physical incivilities in predicting residents’concerns about neighborhood safety. This finding is consistent withresearch on collective efficacy (Sampson et al., 2009), suggestingthat it is really the level of pride in one’s neighborhood, the sense ofneighborhood community, and the willingness to protect one’sneighborhood that predicts concerns about neighborhood safety.When community care and vigilance is high, concerns aboutneighborhood safety are lower (and vice versa). In addition,compared to crimes against person, crimes against property werethe only predictor of concerns about neighborhood safety in thisstudy. This suggests that higher levels of crimes against propertymay directly increase concerns about neighborhood crime.

4.1. Study limitations

Our findings add to the knowledge base on perceptions ofneighborhood crime and safety. However, this study is not withoutlimitations, and our findings should be interpreted withina context of these limitations. First, the cross-sectional nature ofthe data restricts causal interpretations. Second, perceived envi-ronmental crime variables and neighborhood safety concerns werebased on self-reports. Thus, response bias could have affected somefindings, such as actual presence of physical and social incivilities.Conducting an environmental inventory would have helpedcorroborate the overall level of neighborhood physical and socialincivilities. Third, the summed measure for social incivilities hada lower than expected internal consistency reliability. Manystudies document social incivilities through the use of an envi-ronmental inventory. In this study, we documented social incivil-ities through participant self-reports. Future studies shouldexamine the role that data collected through self-reports versusthrough the use of an environmental inventory play in affectinginternal consistency reliabilities. Fourth, our sample was small.Although the sample was representative of the residents whoreceived services from the non-profit organization, it may not havebeen representative of residents living in the larger area. This limitsthe generalizability of our findings. Finally, we measuredcommunity care and vigilance, which is not identical to measuresof collective efficacy and place attachments/territoriality. This mayhave affected some of our findings. Despite these limitations, webelieve that our study yields some novel information, thus, addingto the literature base on this topic.

4.2. Implications for practice

The overarching implication of our findings is that communityworkers/organizers should focus on building residents’ sense ofpride in their neighborhoods and mobilizing residents into actionto improve community conditions, which may directly decreasetheir concerns about neighborhood safety. This might be accom-plished by creatingmore volunteer opportunities for residents of allages. For example, community partnerships can be formed throughsenior centers that serve as volunteer opportunities for youngerresidents (i.e., elementary school-aged, teenagers, young andmiddle-aged adults). In addition, neighborhood watch programscould include residents of all ages. Such efforts could further fostercollective efficacy and place attachments/territoriality, solidifyrelationships, and build mutual trust among residents. This mightresult in residents taking action to reduce neighborhood incivilities,and ultimately, reduce neighborhood crime that affects residents’feeling safe in their neighborhoods. Community collaborations

have been shown to be beneficial to health and well-being ofresidents (Bolda, Lowe, Maddox, & Patnaik, 2005; Sampson,Raudenbush, & Earls, 1999).

Our findings imply that residents’ level of collective efficacy andplace attachment/territoriality may increase when they collectivelytake action to decrease neighborhood incivilities. However, manyagencies are not aware of or fail tomake special efforts to include allresidents in community organizing efforts (Crose & Minear, 1998).There is a need for residents of all ages to get involved incommunity improvement efforts to combat fear of neighborhoodcrime. Our findings indicate that residents who live in vulnerableareas are similar in their concerns about neighborhood safety. Thus,future intervention strategies should be focused on residents of allages getting involved with community-level civic engagementinterventions designed to reduce neighborhood incivilities. This,we believe, will have the greatest potential to bolster residents’perceptions of neighborhood safety.

4.3. Conclusion

Our findings are consistent with research on collective efficacyand place attachment/territoriality in that they show that the senseof pride and the sense of community, as well as residents’ will-ingness to protect their neighborhood, affect how concerned theywill be about neighborhood safety. In addition, our findings areconsistent with research on broken window theory in suggestingthat the presence of any incivility (particularly physical incivility)can also affect residents’ concerns about neighborhood safety.Overall, it does not seem to be an either/or proposition. Rather, allof these variables play important roles in predicting residents’perceptions of, and concerns about, neighborhood safety. Thus,collective efficacy, place attachments/territoriality, and brokenwindow theory are all important predictors of residents’ percep-tions of neighborhood safety.

References

Astor, R., Meyer, H., & Pitner, R. (2001). Elementary and middle school students’perceptions of violence-prone school sub-contexts. The Elementary SchoolJournal, 101(5), 511e528.

Bechtel, R., & Churchman, A. (2002). Handbook of environmental psychology. NewYork: John Wiley & Sons, Inc.

Bolda, E., Lowe, J., Maddox, G., & Patnaik, B. (2005). Community partnerships forolder adults: A case study. Families in Society, 86(3), 411e418.

Brantingham, P. L., & Brantingham, P. J. (1981). Notes on the geometry of crime. InP. L. Brantingham, & P. J. Brantingham (Eds.), Environmental criminology. BeverlyHills, CA: Sage.

Brown, B., Perkins, D., & Brown, G. (2003). Place attachment in a revitalizingneighborhood: Individual and block levels of analysis. Journal of EnvironmentalPsychology, 23(3), 259e271.

Brown, B., Perkins, D., & Brown, G. (2004). Incivilities, place attachments and crime:Block and individual effects. Journal of Environmental Psychology, 24(3),359e371.

Brown, B., Perkins, D., & Douglas, D. (1992). Disruptions in place attachments.Human Behavior & Environment: Advances in Theory & Research, 12, 279e304.

Brown, B., & Werner, C. (1985). Social cohesiveness, territoriality, and holiday-decorations: The influence of cul-de-sacs. Environment and Behavior, 17(5),539e565.

Cisnernos, H. G. (1995). Defensible space: Deterring crime and building community.HUD-1512-PDR. Washington D.C.: U.S. Department of Housing and UrbanDevelopment.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).Hillsdale, NJ: Erlbaum.

Colquhoun, I. (2004). Design out crime: Creating safe and sustainable communities.Boston, MA: Elsevier.

Crose, R., & Minear, M. (1998). Project CARE: A model for establishing neighborhoodcenters to increase access to services by low-income, minority elders. Journal ofGerontological Social Work, 30(3-4), 73e82.

Day, K. (1994). Conceptualizing women’s fear of sexual assault on campus: A reviewof causes and recommendations for change. Environment and Behavior, 26(6),742e765.

Page 7: Making neighborhoods safer: Examining predictors of residents’ concerns about neighborhood safety

R.O. Pitner et al. / Journal of Environmental Psychology 32 (2012) 43e49 49

Eck, J., & Weisburd, D. (1995). Crime places in crime theory. In J. Eck, & D. Weisburd(Eds.), Crime prevention Studies. Crime and place, Vol. 4 (pp. 1034). Monsey, NY:Criminal Justice Press.

Faul, F., Erdfelder, E., Lang, A., & Buchner, A. (2007). G*Power 3: A flexible statisticalpower analysis program for the social, behavioral, and biomedical sciences.Behavior Research Methods, 39(2), 175e191.

Ferguson, K., & Mindel, C. (2007). Modeling fear of crime in Dallas neighborhoods:A test of social capital theory. Crime & Delinquency, 53(2), 322e349.

Fox, J. (1991). Regression diagnostics. Newbury Park, CA: Sage Publications.Franzini, L., Caughy, M., Spears, W., & Esquer, M. (2005). Neighborhood economic

conditions, social processes, and self-rated health in low-income neighbor-hoods in Texas: A multilevel latent variable model. Social Science & Medicine,61(6), 1135e1150.

Goldstein, A. (1994). The ecology of aggression. New York: Plenum Press.Greenberg, A., Rohe, W. M., & Williams, J. R. (1982). Safety in urban neighborhoods:

A comparison of physical characteristics and informal territorial control in highand low crime neighborhoods. Population and Environment, 5(3), 141e165.

Harris, P., & Brown, B. (1996). The home and identity display: Interpreting residentterritoriality from home exteriors. Journal of Environmental Psychology, 16(3),187e203.

Long, A., & Perkins, D. (2007). Community and social place predictors of sense ofcommunity: Amultilevel longitudinal analysis. Journal of Community Psychology,35(5), 563e581.

Newman, O. (1973). Architectural design for crime prevention. (Washington, D.C.:U.S).

Newman, O. (1995). Defensible space: a new physical planning tool for urbanrevitalization. Journal of the American Planning Association, 61(2), 149e155.

Newman, O., & Franck, K. (1980). Factors influencing crime and instability in urbanhousing developments. Washington, DC: US Government Printing Office.

Perkins, D., Florin, P., Rich, R., Wandersman, A., & Chavis, D. (1990). Participation andthe social and physical environment of residential blocks: Crime and commu-nity context. American Journal of Community Psychology, 18(1), 83e115.

Perkins, D., Meeks, J., & Taylor, R. (1992). The physical environment of street blocksand resident perceptions of crime and disorder: Implications for theory andmeasurement. Journal of Environmental Psychology, 12(1), 21e34.

Perkins, D., Wandersman, A., Rich, R., & Taylor, R. (1993). The physical environmentof street crime: Defensible space, territoriality and incivilities. Journal of Envi-ronmental Psychology, 13(4), 29e49.

Pitner, R., & Astor, R. (2008). Children’s reasoning about poverty, physical deterio-ration, danger, and retribution in neighborhood contexts. Journal of Environ-mental Psychology, 28(3), 327e338.

Pitner, R., Yu, M., & Brown, E. (2011). Exploring the dynamics of middle aged andolder adult residents’ perceptions of neighborhood safety. Journal of Geronto-logical Social Work, 54(5), 511e527.

Roman, C., & Chalfin, A. (2008). Fear of walking outdoors: A multilevel ecologicAnalysis of crime and disorder. American Journal of Preventive Medicine, 34,306e312.

Sampson, R. (2004). Neighborhood and community: Collective efficacy andcommunity safety. New Economy, 11(2), 106e113.

Sampson, R., & Graif, C. (2009). Neighborhood social capital as differential socialorganization: Resident and leadership dimensions. American Behavioral Scien-tist, 52(11), 1579e1605.

Sampson, R., Morenoff, J., & Earls, F. (1999). Beyond social capital: Spatial dynamicsof collective efficacy for children. American Sociological Review, 64(5), 633e660.

Sampson, R., & Raudenbush, S. (1999). Systematic social observation of publicspaces: A new look at disorder in urban neighborhoods. American Journal ofSociology, 105(3), 603e651.

SAS Institute Inc.. (2009). Base SAS 9.2 procedures guide. Cary, NC: SAS Institute Inc.Shamard, S. (Winter, 2009). Collective efficacy and fear of crime in the Mat-Su

Borough. Alaska Justice Forum, 25, 4e7.Skogan, W. (1976). Citizen reporting of crime: Some national panel data. Crimi-

nology, 13(4), 535e549.Taylor, R. (1994). Research methods in criminal justice. New York: McGraw-Hill.Taylor, R. (1997). Social order and disorder of street blocks and neighborhoods:

Ecology, microecology and the systemic model of social disorganization. Journalof Research in Crime and Delinquency, 34(1), 113e155.

Taylor, R. (1999). The incivilities thesis: theory, measurement and policy. InR. Langworthy (Ed.), Measuring what matters (pp. 65e88). Washington, DC:National Institute of Justice, Office of Community Oriented Policing Services.

Taylor, R. (2002). Crime prevention through environmental design (CPTED): Yes, no,maybe, unknowable, and all of the above. In R. Bechtel, & A. Churchman (Eds.),Handbook of environmental psychology (pp. 413e426). New York: John Wiley &Sons, Inc.

Taylor, R., & Gottfredson, S. (1986). Environmental design, crime and prevention: Anexamination of community dynamics. In A. J. Reiss, Jr., & M. Tonry (Eds.),Communities and crime (pp. 387e416). Chicago: University of Chicago Press.

Taylor, R., Gottfredson, S., & Brower, S. (1984). Understanding block crime and fear.Journal of Research in Crime and Delinquency, 21(4), 303e331.

Tester, G., Ruel, E., Anderson, A., & Oakley, D. (2011). Sense of place among Atlantapublic housing residents. Journal of Urban Health, 88(3), 436e453.

Wells, W., Schafer, J., Varano, S., & Bynum, T. (2006). Neighborhood residents’production of order: The effects of collective efficacy on responses to neigh-borhood problems. Crime & Delinquency, 52(4), 523e550.

White, G. (1990). Neighborhood permeability and burglary rates. Justice Quarterly,7(1), 57e68.

Wilson, J., & Kelling, G. (1982). Broken windows: The police and neighborhood safety.Retrieved January 16, 2004 from. http://www.theatlantic.com/politics/crime/windows.htm.


Recommended