the built environment, neighborhood crime and constrained physical activity: an exploration of...

Preventive Medicine 47 (2008) 241–251

Contents lists available at ScienceDirect

Preventive Medicine

j ourna l homepage: www.e lsev ie r.com/ locate /ypmed

Review

The built environment, neighborhood crime and constrained physical activity:An exploration of inconsistent findings

Sarah Foster ⁎, Billie Giles-CortiSchool of Population Health, The University of Western Australia, Crawley, Western Australia, Australia

⁎ Corresponding author. Centre for the Built EnviroPopulation Health, The University of Western AustraliaWA 6009, Australia. Fax: +61 8 6488 1199.

E-mail address: [email protected] (S. Fos

0091-7435/$ – see front matter © 2008 Elsevier Inc. Aldoi:10.1016/j.ypmed.2008.03.017

A B S T R A C T
A R T I C L E I N F O
Article history:
Available online 9 April 2008
Keywords:NeighborhoodSafetyCrimeFearPhysical activityWalkingBuilt environment

Objectives. Personal safety is commonly cited in qualitative research as a barrier to local walking, yet therelationship between safety and constrained physical activity has received mixed support in quantitativestudies. This paper reviews the quantitative evidence to date, seeking to explain the inconsistencies, andoffers recommendations for future research.

Methods. A social–ecological framework was adopted to explore the evidence linking crime-relatedsafety, and factors that influence real and perceived safety, with constrained physical activity.

Results. Perceived safety tends to affect the physical activity of groups already known to exhibit greateranxiety about crime; and some elements of the built environment that influence safety appear to constrainphysical activity. However the evidence is somewhat inconsistent, and this may be partly attributed tomeasurement limitations. Many studies employ generic safety measures that make implicit references tocrime or use composite variables that lack specificity. Physical activity outcomes also require consideration,as only activities occurring locally outdoors are likely to be affected by neighborhood crime.

Conclusions. Further research is required to tease out associations between real and perceived crime-related safety and physical activity, ideally employing behaviour and crime-specific measures, and addressingthe moderating role of the social and built environments.

© 2008 Elsevier Inc. All rights reserved.

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241What influences crime and perceptions of safety? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242

Individual characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242Social environmental characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243Built environment characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243Results: the relationship between safety and PA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243

Individual characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243Social environmental characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248Built environment characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248

Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250

nment and Health, School of, 35 Stirling Highway, Crawley

ter).

l rights reserved.

Introduction

While the benefits of physical activity (PA) are well known, there isevidence that activity levels remain inadequate (Centre for DiseaseControl and Prevention, 2001), or are even decreasing in some countries(Armstrong et al., 2000). Globally an estimated 1.9 million deaths areattributed to physical inactivity (World Health Organization, 2002).

mailto:[email protected]

http://dx.doi.org/10.1016/j.ypmed.2008.03.017

http://www.sciencedirect.com/science/journal/00917435

242 S. Foster, B. Giles-Corti / Preventive Medicine 47 (2008) 241–251

Consequently, interventions that increase PA levels could produceconsiderable public health improvements, and neighborhood character-istics that correlate with improved health have become a public healthresearch priority (Owen et al., 2004).

Walking is the most common form of PA (U.S. Department ofHealth and Human Services, 1996), and is usually performed inneighborhood streets and public open spaces (Giles-Corti andDonovan, 2002a). While some neighborhood characteristics havebeen consistently associated with higher levels of walking (Humpelet al., 2004; McCormack et al., 2004; Owen et al., 2004); other featureshave received mixed support in the literature to date. Real orperceived safety from crime is one such characteristic. Qualitativeresearch has repeatedly cited perceptions of crime and safety as abarrier to achieving adequate levels of PA (Miles and Panton, 2006;Eyler et al., 1998; Corti et al., 1996), however the association findsinfrequent support in the quantitative literature.

Research investigating the connection between safety and PAtypically comprises numerous, diverse aspects of safety (e.g., dogs,traffic, infrastructure condition). This paper has a narrower focus;specifically addressing crime-related safety, and other neighborhoodcharacteristics known to influence real or perceived safety from crime.Perceptions of safety can encompass judgments about crime (i.e.,people's estimations of crime and the likelihood that a criminal eventwill occur), or emotional reactions to crime (i.e., fear of crime) (Ferraroand LaGrange, 1987). Ferraro (1995, p.8) defines fear of crime as “anemotional reaction of dread or anxiety to crime or symbols that aperson associates with crime”. To alleviate their fears, people mayalter their behavior according to two patterns: (1) constrainedbehavior, where exposure to potentially dangerous situations isminimised by avoiding certain places and changing behavior; and(2) protective behavior, where security measures are upgraded(Skogan and Maxfield, 1981). Both these responses have implications.It is plausible that people who are fearful about crime may constraintheir PA, particularly if this involves local walking (Ross, 1993).Alternatively, while increased security precautions may alleviate the

Fig. 1. The theore

resident's own fears, alterations to the physical environment canintroduce visual cues thatmay intensify concerns about neighborhoodcrime in other people.

Public health research is increasingly adopting social–ecologicalmodels to integrate the numerous, multi-level correlates influencinghealth-related outcomes (Stokols, 1992). The approach conceives thatenvironments and settings either promote or discourage health-promoting behaviors (Sallis et al., 1998). Accordingly, this paperadopts a social–ecological framework to consider the multiple levelsof influence that affect neighborhood safety and its relationship withPA. The review: (1) summarizes the individual, social and builtenvironment characteristics that influence whether people feel safe;and (2) examines the association between real and perceived crime-related safety, including factors known to influence crime-relatedsafety, and their association with PA. The theorized relationships aredepicted in Fig. 1.

What influences crime and perceptions of safety?

Individual characteristics

It is well documented that certain socio-demographic groups tendto exhibit greater fear of crime (see Hale, 1996). Women and elderlypeople tend to feel more physically vulnerable, hence their greaterconcerns for personal safety. Ethnic minorities and lower socio-economic groups are said to be ecologically vulnerable (Covington andTaylor, 1991), because they have fewer financial resources to protectthemselves or their homes against crime, and often live in neighbor-hoods with concentrated deprivation (Hale, 1996). Direct or indirectencounters with crime can also lead to greater fear of crime, withthose recently victimized typically more fearful. Indirect victimizationentails hearing about a crime second-hand, from friends, family,media or gossip networks. It is far more common than directvictimization and may contribute to the diffusion of fear throughouta community (Skogan and Maxfield, 1981).

tical model.

243S. Foster, B. Giles-Corti / Preventive Medicine 47 (2008) 241–251

Social environmental characteristics

A number of psychosocial processes are associated with perceivedsafety and the incidence of crime. Integration into social networks maybuffer residents from the negative effects of fear and crime by providinga protective environment and reducing feelings of vulnerability (Hale,1996). For example, one study documented that incidental contactbetween neighbors did not impact fear of crime, whereas strongerneighborhood bonds were effective in reducing fear (Riger et al., 1981).However fear of crime can also impinge on a community's ability todevelop social connections, as peoplewhoare fearful and constrain theirbehavior may limit social contact, increasing social isolation andpossibly exaggerating psychological distress (Ross, 1993). To someextent, crime also reflects the social environment: violent and propertycrime has been associated with relative deprivation and low socialcapital (Kawachi et al.,1999); and collective efficacy (trust that residentswill intervene for the community's benefit) has a strong negativeassociation with violence (Sampson et al., 1997).

Built environment characteristics

Good design can promote interaction between residents, facilitatenatural surveillance and encourage a “proprietary attitude toward theneighborhood” (Hale, 1996). The possibility of observation bypedestrian and vehicle traffic or from buildings can increase perceivedsafety and serve to discourage potential offenders (Angel, 1968;Jacobs, 1962; Newman, 1972). In residential areas, natural surveillancecan be promoted through housing design that encourages residents toobserve their streets and minimizes obstacles that restrict visibility(Zelinka and Brennan, 2001).

Theoretically, pedestrians make streets safer, lively and interestingto watch — which encourages surveillance from adjacent buildings(Jacobs,1962). Furthermore, neighborhoods designed to be “walkable”(e.g., accessible destinations, street connectivity) may also generatemore pedestrian traffic, increasing surveillance and influencing safety.However these hypothesized relationships are not necessarilysupported by the evidence. An Australian study comparing a highand low-walkable neighborhood found that perceptions of crime didnot differ (Leslie et al., 2005), and US research revealed objective crimeand walkability were only moderately correlated, with greaterwalkability associated with more crime (Doyle et al., 2006).

Improved lighting has been associated with less crime andincreased pedestrian activity after dark — thereby increasingsurveillance (Painter, 1996). Moreover, enhanced lighting can reducefear of crime (Herbert and Davidson, 1994) and increase respondents'belief that other pedestrians might intervene if required (Vrij andWinkel, 1991). However, while the physical design of a neighborhoodcan provide opportunities for surveillance, the social environmentitself may nonetheless shape the effectiveness of that surveillance.British researchers found superior natural surveillance was onlymarginally protective against vandalism, after accounting for proxi-mity to council housing. They concluded that the social composition ofthe surrounding population was more influential than surveillanceper se (Mayhew et al., 1979).

Other elements of the physical environment provide negativevisual cues that potentially amplify residents' perceptions of crimeand heighten fears. The presence of physical disorder or “incivilities”(e.g., vandalism, litter, graffiti) (Skogan and Maxfield, 1981) isassociated with increased burglary (DeFrances and Titus, 1993),greater perceived crime and fear of crime (Lewis and Maxfield,1980), even when reported crimes are low (Sooman and Macintyre,1995). Moreover, when incivilities are not remedied quickly, residentsmay interpret neighborhood decline as being worse than it really is(Skogan, 1990). Incivilities appear to act as an intermediate variable,where the presence of disorder increases concern about neighborhoodcrime and generates more fear (Lewis and Maxfield, 1980).

House and garden maintenance can also influence crime andperceived safety. Well maintained housing signals the resident'sproprietary, and offers a symbolic distinction between public andprivate space. Research has associated unkempt lawns and litter withcrime vulnerability in the suburbs (Brown et al., 2004). Moreover, thecondition and maintenance of suburban neighborhoods is associatedwith perceived safety (Austin et al., 2002; Wood et al., 2008). In cityneighborhoods, it has been postulated that obvious security measuresin public space (e.g., observation cameras) may signal to pedestriansthat the area is unsafe (Painter, 1996). If this notion is accurate; visiblehome security precautions may also generate concerns about safety,and may represent another form of incivility, particularly in suburbanneighborhoods.

Methods

Electronic databases, including PsycINFO, Medline, Web of Science, Science Direct,ProQuest Social Science Journals and Pubmed were searched using combinations of thekey words: neighborhood, walking, PA, exercise, crime, safety and fear. Only peerreviewed, English language, quantitative papers, published before July 2007 thataddressed some form of crime-related safety and a PA outcome in adult populationswere reviewed.

Results: the relationship between safety and PA

Forty-one papers met the search criteria. Table 1 summarizes theresults, including crime-related safety measures, or neighborhoodcharacteristics that influence crime-related safety; and any significantassociations with PA. Some studies do not explicitly mention crimedue to a reliance on global measures of safety, which do not openlystate the source of insecurity. Nonetheless, global measures of safetyhave generally been adopted to gauge perceptions of crime-relatedsafety, and are there-fore highly relevant to the review.

Individual characteristics

As discussed, certain groups in society tend to feel more vulnerableto crime and this potentially confounds the relationship betweensafety and PA. There is evidence that physical inactivity is higheramong those who perceive their neighborhood to be unsafe fromcrime (Weinstein et al., 1999). Indeed, feeling unsafe was associatedwith inactivity in women, older adults, ethnic minorities and thosewith lower educational attainments. Moreover, after adjusting for raceand education, older adults (N65 years) who felt unsafe were morethan twice as likely as younger adults (V64 years) to be inactiveHowever, most research does not demonstrate the associationbetween individual characteristics, safety and PA so elegantly.

Several studies focused exclusively on women (King et al., 2000;Wilcox et al., 2000; Wilcox et al., 2003; Ball et al., 2007) and minoritywomen (Ainsworth et al., 2003; Evenson et al., 2003; Hooker et al.,2005; Voorhees and Young, 2003; Wilbur et al., 2003a; Wilbur et al.,2003b; Young and Voorhees, 2003). While most found no association,two studies found significant associations between crime-relatedsafety and PA (Wilbur et al., 2003a; Wilcox et al., 2003). In an areanoted for high homicide rates, African-American women were 2.4times more likely to do any activity if they perceived theirneighborhood to be extremely or somewhat safe from crime (Wilburet al., 2003a); and for a sample of rural white and African-Americanwomen, feeling safe was associated with higher levels of PA (Wilcoxet al., 2003). While minority womenmight experience greater anxietyabout crime-related safety, this is not necessarily reflected inconstrained PA among these samples. Notably, in a sample ofAfrican-American and white adults (men and women), perceivingthe neighborhood to be safe from crimewas associated with increasedodds of walking for white people, whereas there was no apparentassociation among African-Americans (Hooker et al., 2005).

Studies examining the effects of safety on PAwith adult populationsoften identify different effects for men and women (Suminski et al.,

Table 1Characteristics of studies examining relationships between neighborhood safety variables and physical activity

Reference Number, age,gender

Setting Safety, social and built environmentmeasures

PA outcome Statisticaladjustment

Significant associations betweensafety items and PA outcome

Ball et al.(2007)

N=1282.18–65 years,F

Melbourne,Australia,Suburban

Subjective: Composite safety variableincludes: (1) my neighborhood issafe for walking; (2) it is safe outwalking day or night; (3) the streetsare well lit at night

Walking for leisure; walking fortransport

E, L, SS, DO,CM, Env

High neighborhood safety associatedwith more leisure time walking; butnon significant when personalmediators were added to the model

Mota et al.(2007)

N=181,M:M age=76.6,F:M age=79.1

Porto,Portugal,Urban

Subjective:Neighborhood safetyincludes: the crime rate in myneighborhood makes it unsafe orunpleasant towalk inmyneighborhood

Work activity; sports activity;leisure time activity; global activity

Stratified by S Concerns about neighborhood safetyassociated with less total activityand less leisure activity

Harrison et al.(2007)

N=15461,M=49.8 years,M+F

2 districts inNorthwestEngland,Setting NS

Subjective: In their neighborhoodhow much of a problem wasvandalism; assaults and muggingsand speeding traffic. Had they beenthe subject of a personal crime in thelast year? Whether they felt safe outand about in their neighborhood:(1) during the day and (2) duringthe night. Sense of belonging

Physically active A, S, Eth, D Peoplewho felt safe “during the day”and “during the night” were morelikely to be physically active

Shenassa et al.(2006)

N=5338,18–65 years,M+F

8 Europeancities, Urban/suburban/rural

Subjective: Do you feel safe returningto your home when it is dark?Objective (environmental audits)Dwelling and immediateenvironment rated for amount oflitter and presence of graffiti

No current exercise; occasionalexercise; frequent exercise

A, S, E, MS,DS, C, L,stratified by S

Safety associated with increasedodds of occasional exercise (men andwomen); occasional exercise (menonly); and frequent exercise (womenonly). Litter associated with reducedodds of occasional exercise andfrequent exercise (men and women),and frequent exercise (men only)

Piro et al.(2006)

N=3499,75–76 years,M+F

Oslo, Norway,Setting NS

Subjective: Would you feel safewalking alone in your neighborhoodin the evening?Objective: Registered cases ofviolence per 1000 inhabitants inOslo's 25 administrative areas

Less than 1 h PA/week; more than1 h PA/week

MS, I, E, R,MC, stratifiedby S

For men, high neighborhood levelviolence (objective) was associatedwith less PA; For women, low safety(subjective) was associated with lessPA

Pikora et al.(2006)

N=1678,18–59 years,M+F

Perth,Australia,Suburban

Objective (environmental audits)Personal safety: (1) lighting overpath; (2) surveillance from housesAesthetics (streetscape) includes:(1) garden maintenance; (2) streetmaintenance; (3) cleanliness

Walking near home for recreation;walking near home for transport

Demographic,social andindividualfactors, areaSES, Env

No association between objectivesafety features (either mergedvariable or its elements) and walking

Doyle et al.(2006)

N=9252,18+ years,M+F

Counties inlarge urbanareas, USAUrban

Objective: County crime rates forserious crimes known to police,including: murder, non-negligentmanslaughter, forcible rape, robbery,aggravated assault, burglary,larceny-theft andmotor vehicle theft

Frequency of walking: walking onemile or more without stopping inthe last month

A, S, Eth, I, E,SS, Smoke, R,walkability

Crime was associated with lesswalking (non-significant), butsignificantly associated with higherBMI

Vest andValadez(2005)

N=1635,Adults,M+F

Austin/TravisCounty, Texas,USA SettingNS

Subjective: How safe from crime doyou consider your neighborhood tobe? For walking at night, would youdescribe the street lighting in yourneighborhood as very good/good;fair; poor/very poor?

Leisure time physical inactivity A, S, E, Eth Lower safety from crime associatedwith increased odds of physicalinactivity

van Lentheet al. (2005)

N=8767,20–69 years,M+F

Eindoven,Netherlands,Urban

Objective: Neighborhood safety:Amount of police attention requiredin an area as rated by municipalservices employees:(1) neighborhood requires veryspecial attention regularly; (2) areathat requires attention; (3) area thatcan be controlled with normalattention; (4) quiet area with veryincidental problems of externalcauses; (5) very quiet area withhardly any conflicts

Almost never walking or cycling toshops or work; almost neverwalking, cycling and gardening inleisure time; almost neverparticipating in sports

A, S, E, SES More police attention associatedwith almost no sports participationThe relationship betweenneighborhood SES and almost neverparticipating in sports activities waspartly mediated by police attention

Suminski et al.(2005)

N=474,18+ years,M+F

Midwesternmetropolitanarea, USASetting NS

Subjective: Composite safetymeasure includes 4 items: Howwould you describe the: (1) amountof traffic; (2) speed of traffic;(3) lighting; (4) level of crimein your neighborhood?

Walking for exercise; dog walking;walking for transport; engage inactivity; engage in activity in theirneighborhood

A, E, stratifiedby S

For females higher safety (middletertile compared to lowest tertile)was associated with increased oddsof walking for exercise and dogwalking

Li et al. (2005) N=577,65–94 years,M+F

Portland,Oregon, USASetting NS

Subjective: Composite “safety forwalking” score includes: (1) it is safeto walk or jog alone in myneighborhood during the day;(2) unsafe sidewalks (obstacles towalking) are a problem

Walking, strolling and other PAaveraged into a walking activityscore

A, S, MS, E, I,HS, Env

At resident level, higher “safety forwalking” score associated withwalking


Table 1 (continued)





Lim and Taylor(2005)


NSW,Australia,Urban/rural

Subjective: Whether respondentfeels safe in the neighborhood

Adequate PA; inadequate PA A, S, L, E, Em,K6, PF, MC,Diet, Care, IT,Lang, LA, Fall

No significant association betweencrime-related safety items and PA

Hooker et al.(2005)

N=1165,18–96 yearsM+F AfricanAmerican/white

SouthCarolina, USARural

Subjective: Streetlight quality inneighborhood (good, fair, poor);neighborhood safe from crime (safe,not safe); public recreation facilitiesare safe (yes, no)

Walking 150+ min per week; notwalking 150 min per week

A, S, Estratified byEth

Whites who perceived theirneighbourhoods as safe from crimewere more likely to walk

Hoehner et al.(2005)

N=1068,18–96 years,M+F

St Louis andSavannah,USA Urban

Subjective: Aesthetics: (1) myneighborhood is generally free fromgarbage, litter, broken glass; (2) myneighborhood is well maintained.Social environment: How safe fromcrime do you feel while you arewalking or riding your bike in yourneighborhood?Objective: (environmental audits)Traffic safety (sum of 7 items,includes street lighting); aesthetics(% segments with no or littlegarbage, litter or broken glass);physical disorder (sum of 8 items:beer/liquor bottles/cans, cigarette/cigar butts/packages, condoms, drugrelated paraphernalia, garbage, litteror broken glass, abandoned cars,graffiti and broken windows); crimewatch signs (count of neighborhoodor crime watch signs)

Any transportation activity(walking or bicycling); metrecommendations solely throughtransportation activity; metrecommendations solely throughrecreational activity

A, S, E, Perceiving neighborhood free ofgarbage etc. was associated withlower odds of meetingrecommendations (transportactivity); Perceiving neighborhoodto be well maintained associatedwith lower odds of meetingrecommendations (transportactivity); Less objective garbage wasassociated with lower odds of anytransportation activity/meetingrecommendations; Less objectivephysical disorder associated withlower odds of any transportationactivity/meeting recommendations

Ellaway et al.(2005)

N=6919,Adults,M+F

8 Europeancountries,Setting NS

Objective: Trained surveyors ratedresidential neighborhoods for:(1) litter, graffiti and dog mess;(2) vegetation and greenery

PA in two levels: Seldom/never;often

A, S, SES, L Neighborhoods with high levels ofincivilities were associated withlower odds of PA

Duncan andMummery(2005)


Rockhampton,Queensland,AustraliaSetting NS

Subjective: Environmental attributesinclude: It is safe to walk in theneighborhood; crime is high in theneighborhood; your neighborhood iskept clean and tidy; the streets arewell lit

Sufficient PA; any recreationalwalking.

A, I, S, BMI, SS,SE

Not agreeing the neighborhood wasclean and tidy associated withincreased odds of sufficient PA

Humpel et al.(2004)

N=399,40+ years,M+F

Regionalcoastal cityand suburbs,AustraliaUrban/suburban

Subjective: Composite safety variableincludes: Crossing busy roads is a bigproblem/crossing busy roads is not aproblem; traffic makes it dangerousor unpleasant/traffic is not aproblem; I don't feel safe at all/I feelvery safe; dogs always put me offwalking/dogs never put me offwalking

Walking in and aroundneighborhood: neighborhoodwalking; exercise walking;pleasure walking; walking to get toand from places

A, E, stratifiedby S

Males with high perceived safety(compositemeasure) were less likelyto walk for pleasure

Foster et al.(2004)

N=4265,16–74 years,M+F

England,Setting NS

Subjective: It is safe to go out walkingduring the day on your own; it is safeto go our walking during the nighton your own

Walking for at least 15 min/week;walking for at least 150 min/week

A, SES, E, HS,car use,stratified by S

Women who felt it was unsafe to goout walking (during the day) wereless likely to walk for 15 min/week

Young andVoorhees(2003)

N=234,20–50 years,F AfricanAmerican

Baltimore,Maryland,USA Urban

Subjective: For walking at night,would you describe the streetlighting where you live as very good,good, fair, poor, very poor?How safe from crime is itwhere you live? Would you sayextremely safe, somewhat safe,slightly safe, not at all safe? Sense ofcommunity

Met recommendations; insufficientactivity; inactive

A, I, E, MS, C,HS(unadjustedreported)


Wilcox et al.(2003)

N=102,50+ yearsF AfricanAmerican andwhite


Subjective: Street lighting in theirneighborhood; Safety of theneighborhood during the day, atnight, overall (average used for a“neighborhood safety” rating)

Summary score for total PA fromleisure activities, housework, paidor volunteer work over the past7 days

A, PDB, DSy,HCP, nosidewalks

Increased perceived safety wasassociated with higher levels of PA

Wilbur et al.(2003)


6 Chicagocommunities,USA Urban

Subjective: For walking at night,would you describe the streetlighting where you live as very good,good, fair, poor, very poor? How safefrom crime is it where you live?Would you say extremely safe,somewhat safe, slightly safe, not atall safe? Sense of community

Meets recommendations; anyactivity; no activity

A, E, I, MS, C,HS(unadjustedreported)

Women were more likely to beactive if they perceived theirneighbourhoods to be extremely orsomewhat safe from crime

(continued on next page)(continued on next page)

Table 1 (continued)


Table 1 (continued)





Wilbur et al.(2003)

N=300,20–50 years,F Latinas

3 Chicagocommunities,USA Urban


Meets recommendations; anyactivity; no activity

A, E, I, MS, C,HS, Ac,(unadjustedreported)


Voorhees andYoung (2003)

N=285,20–50 years,FHispanic/Latina

Arlington andFairfax inVirginia, USASetting NS



A, E, C, HS No significant association betweencrime-related safety items and PA

Troped et al.(2003)

N=413,34–68 years,M+F

Arlington,MassachusettsUSA Suburban

Subjective: Neighborhoodcharacteristics include: (lack of) highcrime rates; street lightsNeighborhood safety: how safe doyou feel walking in yourneighbourhood during the day?

Recreational activity; transpor-tation PA (time going to/fromwork,school or store)

A, SE Street lights were associated withincreased transport PA, but nonsignificant when otherenvironmental variables wereentered into the model

Parks et al.(2003)


USA Urban/suburban/rural

Subjective: Personal barriers include:no safe place


A, S, Eth, E,stratified by U

No significant association between“no safe place” and PA

Huston et al.(2003)


6 counties inNorthCarolina, USASetting NS

Subjective: Neighborhoodcharacteristics include: presence ofstreet lights. Safety from crime: howsafe from crime do you consider yourneighborhood to be?

Any activity; recommendedactivity

A, S, E, Eth Presence of streetlights associatedwith any activity, but non significantafter adjustment

Eyler et al.(2003)


USA, Urban/suburban/rural

Subjective: Potential barriersinclude: no safe place; high crime;no street lights

Regular walkers; occasionalwalkers; never walkers

A, E, Eth “No safe place” associated withgreater odds of never walking

Evenson et al.(2003)

N=671,20–50 years,F Latinas

NorthCarolina, USASetting NS



A, HS, C, MS,E, Ac

Item from sense of community scale,“agreeing your neighbors could becounted on to help if someone wasdestructive to property in yourneighborhood” was associated withany activity

DeBourdeaudhuijet al. (2003)

N=521,18–65 years,M+F

Ghent,BelgiumUrban/suburban/rural

Subjective: Perceived safety fromcrime score: the crime rate in myneighborhood makes it unsafe to goon walks during the day and duringat night. Neighborhood aestheticsscore includes: my neighborhood isgenerally free from litter

Sitting; walking; moderateintensity (walking not included)and vigorous intensity activities

A, E, Em,stratified by S

Greater perceived safety from crimewas associated with less time spentsitting for males

Ainsworthet al. (2003)





E, L No significant association betweencrime-related safety items and PA

Giles-Cortiand Donovan(2002)

N=180318–59 years,M+F

Perth,AustraliaSuburban

Subjective: Composite neighborhoodattractiveness, safety and interestscore includes: the neighborhood iswell maintained; the neighborhoodis safe for walking; it is safe outwalking day or night. Separate item:whether the streets are well lit

Sufficiently active; walking asrecommended; exercising atvigorous levels; any walking fortransport; any vigorous activity

A, S, C, E, I, WS Composite attractiveness, safety andinterest score associated withgreater odds of walking forrecreation, walking asrecommended, any vigorous activity,and exercising vigorously at therecommended level

Craig et al.(2002)

N=27neighborhoods

Ontario,Quebec andAlberta,Canada,Urban/smallurban/suburban

Objective: Composite environmentalscore includes: “safety from crime”(lighting, front porches; escaperoutes; number of people around,potential avenues to escape;property maintenance; street typeetc.); “potential for crime” (graffiti,vandalism, disrepair, potential forlurking, street lighting, propertymaintenance); and other variablesincluding: traffic threats,destinations, walking routes,aesthetics and social dynamics

Walking to work (From 1996Census)

U, E, I, P Increase in environmental scoreassociated with increased walking towork

Table 1 (continued)


Table 1 (continued)





Carnegie et al.(2002)

N=1200,40–60 years,M+F

Wollongong,Australia,Setting NS

Subjective: Perceived safety ofwalking in the area during the dayand night; composite aestheticenvironment score includes: safetyfor walking (day)

PA: active/inactive. Walking:0–20 min/week; 21–120 min/week;N2 h/week

A, S, E Negative perceptions of aestheticenvironment (including safetywalking during the day) associatedwith less walking

Steptoe andFeldman(2001)

N=658,18–94 years,M+F

London area,Urban

Subjective: Extent of problems inyour area including: litter in thestreets; smells and fumes; walkingaround after dark; problems withdogs; noise from traffic and otherhomes; lack of entertainment; trafficand road safety; places to shop;vandalism; disturbances byneighbors or youngsters. Socialcohesion and informal social control

Vigorous PA: classified as active orinactive

A, S, SES, D No association betweenneighborhood problems and PA, butneighborhood problems wereassociated with poorer self-ratedhealth, psychological distress andimpaired physical function

Brownsonet al. (2001)


USA SettingNS

Subjective: Presence or absence ofneighborhood characteristicsincluding: high crime; no safe place;heavy traffic; unattended dogs;street lighting

Moderate; insufficient; inactive A, S, Eth, I, E High crime and “no safe place”associated with less PA, but nonsignificant after adjustment

Wilcox et al.(2000)

N=2338,40+ years,F

USA Urban/rural

Subjective: Psychosocial barriers,includes “lack a safe place”Environmental factors include:heavy traffic; street lights;unattended dogs; high levels ofcrime

Sedentary; active; underactive A, E, Eth,geographicalregion,stratified by U


Ross (2000) N=2482,18–92 years,M+F

Illinois USAUrban/suburban/small city/town/rural

Subjective: Number of days last weekthat someone: (1) feared beingrobbed, attacked or physicallyinjured; (2) worried that their homewould be broken into; (3) felt afraidto leave the house

Walking (days/week); exercising(days/week)

A, S, MS, Eth,E, I, SES, L

Increased fear associated withreduction in walking

King et al.(2000)

N=2912,40+ years,F

USA Urban/rural/other

Subjective: Lacking safe place toexercise; how safe is it to walk or jogin their neighborhood during theday? Were the following present:heavy traffic, street lights,unattended dogs, high crime

Sedentary; under active; active A, E, Em, MS,L, stratified byEth


Booth et al.(2000)

N=449,60+ years,M+F

Australia,Excludessparelypopulatedareas

Subjective: Safety or difficulty ofwalking in the neighborhood duringthe day; footpaths perceived as safefor walking

Inactive; active A, S Perceiving footpaths as safe forwalking was associated with beingactive

Weinsteinet al. (1999)

N=12 767,18+ years,M+F

Various states,USA SettingNS

Subjective: How safe from crime doyou consider your neighborhood tobe?

Physically inactive Eth, E Less perceived safety from crimeassociated with physical inactivity

Sallis et al.(1997)

N=110,M=20.6 years,M+F

USA, Urban Subjective: Compositeenvironmental score includes: streetlights, (lack of) high crime rates andperceived safety (how safe you feelwalking in your neighborhoodduring the day?)

Strength exercise; vigorousexercise; walking for exercise

A, S, Eth, SES No significant association betweencomposite environmental score(including crime-related safetyitems) and PA

A, age; Ac, acculturation; BMI, body mass index; C, number of children; Care, care giving; CM, club membership; D, deprivation; Diet, adequate intake of fruit and vegetables; DO, dogownership; DS, disability status; DSy, depressive symptoms; E, education; Em, employment; Env, environmental variables; Eth, ethnicity; F, female; Fall, afraid of fall; HCP, health careprovider; HS, health status, I, income; IT, independent travel; K6, psychological distress measure; L, location; LA, living arrangements; Lang, language spoken at home; M, male; MC,medical conditions; MS, marital status; P, poverty; PDB, positive decision balance, PF, physical functioning; R, length of residence; S, sex; SE, self-efficacy; SES, socio-economic status;Setting NS, study setting not specified as urban, suburban or rural; Smoke: smoking history; SS, social support; U, urbanization; Unadjusted reported, statistical adjustment madelittle difference to the results so unadjusted figures are reported; WS, work status.All study designs are cross-sectional and all PA outcomes are self-reported. Only crime-related safety measures, or neighborhood features that might influence crime-related safety,are included in the table. Some studies document significant findings between other elements of safety (e.g., traffic, dogs) and PA; however these findings have not been included.

Table 1 (continued)


2005; Shenassa et al., 2006; Foster et al., 2004; Humpel et al., 2004; DeBourdeaudhuij et al., 2003), with women more likely to constrain theiractivity. Suminski et al. (2005) found perceived safety – an aggregatemeasure of respondent's ratings for traffic volume and speed, lightingand crime – was a significant correlate for female walking, but not formales. They reported that women were over 4.5 times more likely towalk for exercise, or 3.3 times more likely to walk their dog if theyperceived average levels of safety, compared with below average levelsof safety in the neighborhood. Although the findings were significant,the approach taken to measure safety – which combined differentaspects of safety – obscures the true source of residents' concerns.

Other research with adults highlights that safety concerns canrestrict PA across both sexes (Vest and Valadez, 2005; Harrison et al.,2007; Shenassa et al., 2006; Eyler et al., 2003), but further studiesreported no significant associations (Carnegie et al., 2002; Duncan andMummery, 2005; Hoehner et al., 2005; Huston et al., 2003; Sallis et al.,1997; Troped et al., 2003; van Lenthe et al., 2005). However, generallyinsecurity about crime and safety is more prevalent among women(Carnegie et al., 2002), and these concerns can manifest themselves inless PA.

Research undertaken with older populations also found mixedresults. Some studies supported the connection between crime-


related safety and reduced PA (Wilcox et al., 2003; Piro et al., 2006; Liet al., 2005; Mota et al., 2007), but others found no association (Limand Taylor, 2005; Booth et al., 2000). In one example, Piro et al. (2006)examined subjective safety and objective crime in a sample of elderlynon-institutionalised residents. They found objective neighborhood-level violence was significantly associated with lower odds of PAamong men, whereas less perceived safety was associated with lowerodds of PA among women. While violence was clustered in particularneighborhoods, the authors noted that low perceived safety amongwomen was not clustered; indicating that perceptions of safety aremore psychologically complex and independent of reported crime.These results support the notion that concerns about safety mightconstrain PA in older populations, but also highlights the potential forgender to influence perceived safety and restrict activity. Furthermore,these findings underscore that safety, measured in different ways maydifferentially influence sub-populations, affecting research outcomes.

Several studies considered differential effects for lower socio-economic groups (Brownson et al., 2001; Parks et al., 2003; Ross,2000; van Lenthe et al., 2005; Wilson et al., 2004). Residents in poorareasweremore fearful of criminal victimization; however respondentsin these neighborhoods also tend to walk more than those in lessdisadvantaged neighborhoods (Ross, 2000). Ross (2000) suggests thatthis finding may be explained by the normative behaviors of theneighborhoods studied. Disadvantaged neighborhoods in the US tend tohave higher residential densities thatmay encouragewalking, and therecouldbea “contagioneffect” that promotes a culture of socializingon thesidewalks (p.272). Other research found that lower income groups hadhigher self-reported exposure to crime (Brownson et al., 2001), andperceived less safety from crime (Wilson et al., 2004). However, in thelatter study, different perceptions of safety were not supported by thecorresponding objective measures. Nonetheless, Brownson et al. (2001)found only amodest inverse association between the perception of highcrime and PA, which was non-significant after multivariate adjustment.

Social environmental characteristics

Fewstudies examining the safetyand PA relationship include a socialenvironmental measure. Those incorporating sense of belonging(Harrison et al., 2007) and sense of community (Ainsworth et al.,2003; Evenson et al., 2003; Voorhees and Young, 2003; Wilbur et al.,2003a; Wilbur et al., 2003b; Young and Voorhees, 2003) found nosignificant association with PA. However, research adopting measuresbetter conceptualized as collective efficacy did identify an association.Evenson et al. (2003) note that one component of their scale: “agreeingyour neighbors could be counted on to help if someonewas destructiveto property inyour neighborhood”was significantly associatedwith anyactivity. Collective efficacy may encourage walkers by alleviatingconcerns (i.e., residents will intervene if required); whereas other socialconstructs may be an outcome of pedestrian interaction.

Built environment characteristics

Two studies objectively measured the potential for surveillance andwalking outcomes (Pikora et al., 2006; Craig et al., 2002). Craig et al.(2002) considered various street features that promote surveillance andcrime vulnerability. Auditors classified numerous elements into themes(e.g., “safety from crime” included lighting, front porches, escape routes,maintenance etc.). Ultimately, the themes, “potential for crime” and“safety from crime”were found to significantly contribute to an overallenvironmental score; which was subsequently associated withincreased walking to work in the study neighborhoods (Craig et al.,2002). Alternatively, Pikora et al. (2006) found no association betweensurveillance from housing and walking.

Half the papers investigating safety and PA included some measureof street lighting (Ainsworth et al., 2003; Ball et al., 2007; Brownsonet al., 2001; Craig et al., 2002; Duncan and Mummery, 2005; Evenson

et al., 2003; Eyler et al., 2003; Giles-Corti and Donovan, 2002b; Hustonet al., 2003; King et al., 2000; Pikora et al., 2006; Sallis et al., 1997;Suminski et al., 2005; Troped et al., 2003; Vest and Valadez, 2005;Voorhees and Young, 2003; Wilbur et al., 2003a; Wilbur et al., 2003b;Wilcox et al., 2003; Wilcox et al., 2000; Young and Voorhees, 2003)although only one study of suburban residents identified a significantassociation between street lighting and PA, after controlling for age andself-efficacy (Troped et al., 2003). Two studies merged street lightinginto a combined safety (Humpel et al., 2004; Suminski et al., 2005) orenvironmental levelmeasure (Craig et al., 2002), andwhile these studiesalso identified significant associations with walking, the independentcontribution of lighting remains unclear.

Other elements of the built environment provide visual cues thatcan influence safety perceptions. Several studies reviewed includeddisorder as an independent variable, and examined its connectionwith PA (Ellaway et al., 2005; Harrison et al., 2007; Hoehner et al.,2005; Shenassa et al., 2006; Steptoe and Feldman, 2001). Europeanresidents in neighborhoods with high levels of objectively rateddisorder were about 50% less likely to be physically active and about50% more likely to be overweight or obese (Ellaway et al., 2005).Hoehner et al. (2005) also examined the role of incivilities but withconflicting results. For both perceived and objectively rated disorder,lower levels of disorder were associated with reduced odds oftransportation walking. Like traffic exposure, more frequent walkersmay be exposed to, and therefore aware of, higher levels of disorder(Duncan and Mummery, 2005). However, higher levels of physicaldisorder tend to cluster in denser urban areas, which have more non-residential land uses (Taylor et al., 1995) suggesting that the studycontext (i.e., urban, suburban, rural) and neighborhood walkabilitymay confound the relationship between disorder and PA. Further-more, individual and neighborhood level socio-economic status (SES)complicate the association, as incivilities concentrate in low SESneighborhoods (Steptoe and Feldman, 2001) and residents in theseareas oftenwalk out of necessity (Ross, 2000). These findings highlightthe need for consistent adjustment for potential confoundingvariables.

With regard to neighborhood maintenance, objective ratings werenot associated with walking (Pikora et al., 2006); however studiesusing subjectivemeasures hadmixed findings. Duncan andMummery(2005) identified that people who perceived lower levels ofneighborhood maintenance were 3.13 times more likely to achievesufficient levels of PA. Similarly, Hoehner et al. (2005) linkedperceptions of a well maintained neighborhood with less transporta-tion activity. Again, those who frequent the neighborhood publicspaces may be more conscious of their presentation (Duncan andMummery, 2005). However, a composite measure (including percep-tions of safety and maintenance) was linked with increased PA (Giles-Corti and Donovan, 2002b). Despite conflicting evidence, neighbor-hoodmaintenancewarrants attention as a feature that might promoteor detract from perceived safety, and influence PA.

Discussion

Studies exploring neighborhood safety and PA have identifiedsomewhat mixed results. The research has a number of limitationsassociated with the measurement of crime-related safety. Many studiesrely on “global”measures of safety, which do not explicitly mention thesource of insecurity, and may be interpreted in various ways byrespondents (Ball et al., 2007; Booth et al., 2000; Carnegie et al., 2002;Foster et al., 2004; Giles-Corti and Donovan, 2002b; Hoehner et al.,2005; Humpel et al., 2004; Li et al., 2005; Lim and Taylor, 2005; Parkset al., 2003; Piro et al., 2006; Shenassa et al., 2006;Wilcoxet al., 2003). Inquestions such as, “how safe do you feel walking in your neighborhoodat night” (Piro et al., 2006), or “do you feel safe returning to your homewhen it is dark?” (Shenassa et al., 2006), the threat is implied but thesource of insecurity is not clearly stated (i.e., is it crime, dogs, or traffic


that causes the respondent to feel unsafe?) (Hale, 1996). Moreover,respondents who feel unsafe may not actually be fearful, but may judgethe risk of victimization to be high (Ferraro and LaGrange, 1987). Globalmeasures of safety have been criticized for overestimating concernabout crime as they present respondents with a situation that manywould rarely encounter, but nonetheless feel apprehensive about (i.e.,being alone, walking and at night) (Ferraro and LaGrange, 1987).

Furthermore, many safety ratings are based on judgments aboutcrime and disorder (Ainsworth et al., 2003; De Bourdeaudhuij et al.,2003; Evenson et al., 2003; Eyler et al., 2003; Hooker et al., 2005;Huston et al., 2003; Mota et al., 2007; Steptoe and Feldman, 2001;Suminski et al., 2005; Vest and Valadez, 2005; Voorhees and Young,2003;Weinstein et al., 1999;Wilbur et al., 2003a;Wilbur et al., 2003b;Young and Voorhees, 2003), whereas it is plausible that emotionalresponses might elicit more consistent associations between “neigh-borhood safety” and PA. For example, while respondents may believecrime is a problem in their neighborhood, this perception may notcorrelate with constrained PA unless those respondents are alsofearful about the crime they perceive. Ferraro and LaGrange (1987)recommend that researchers measuring fear of crime include itemsthat: (1) refer explicitly to “fear” rather than worry or concern aboutcrime; (2) specifically mention crime; and (3) are not hypothetical.Only one author reviewed here met these criteria: Ross (2000)identified that people who fear victimization are less likely to walk.

Several studies merged diverse aspects of safety into a compositevariable (Ball et al., 2007; Carnegie et al., 2002; Giles-Corti andDonovan, 2002b; Humpel et al., 2004; Li et al., 2005; Sallis et al., 1997;Suminski et al., 2005). Sallis et al. (1997) combined neighborhoodfeatures, perceived safety and neighborhood character into anenvironmental scale and found no association with PA. Wherecomposite measures are analysed, the combined measure mayobscure effects which might have been apparent had the analysisbeen conducted on component variables (Humpel et al., 2002).Another study merged resident perceptions of traffic, lighting andcrime (Suminski et al., 2005). While this study identified significantassociations between safety and walking, the effort may be mean-ingless if the authors cannot identify which dimension of safety (e.g.,crime or traffic) is at the core of the problem, and where to channelappropriate interventions.

Research has adopted subjective or objective assessments of safety,and some has incorporated both. Objective measures includedreported crime (Piro et al., 2006; Doyle et al., 2006) and rating policeattention (van Lenthe et al., 2005). It is well documented that crimestatistics and fear of crime are only weakly associated (Lewis andMaxfield, 1980), and evidence suggests that many more people areafraid of crime than are actually victimised (Hale, 1996). As such,perceptions of crime may have a more powerful effect on behaviorthan objective measures. Other studies used auditors to assess aspectsof the environment that promote or deter feelings of safety, includingdisorder, street lighting and surveillance (Ellaway et al., 2005;Shenassa et al., 2006; Craig et al., 2002; Pikora et al., 2006). However,the majority of the research tackling the association between safetyand PA relies on subjective data. Pikora et al. (2006) comment thatperceptions of safety may be a more important predictor of walkingthan objective safety ratings, and futurework should incorporate bothaspects. Moreover, research that combines subjective perceptions andobjective neighborhood features may provide greater insight as towhether neighborhood-level interventions have the potential to easesafety concerns.

Physical environmental factors also contribute to the relationshipbetween crime-related safety and PA. For example, street lighting andsurveillance from housing provide the opportunity for people tomonitor the neighborhood; and incivilities present visual cues thatcan amplify residents' fears about crime. While these neighborhoodfeatures may not directly affect PA, they may affect perceived safety.This review has focused on the safety-PA relationship; hence it does

not provide a detailed exploration of the correlates of safety.Nevertheless, it is worth considering the theoretical pathway linkingfeatures of the built environment with PA, as perceived safety maymediate any direct association. Similarly, the social environment andits influence on PA may be mediated by crime and perceived safety,and this warrants attention.

Variations in study locations may also contribute to mixedfindings. The literature spans diverse climates, locations and cultures,which may promote or inhibit outdoor activity; and the degree ofurbanization (i.e., urban, suburban, rural) may affect exposure tofactors that influence safety perceptions (e.g., crime, incivilities,surveillance, lighting) (Wilcox et al., 2000). Moreover, studies insingle settings may have limited variability in objective or perceivedsafety, thereby hindering the possibility of finding significant associa-tions between safety and PA (Hoehner et al., 2005).

Finally, walking or PA outcomes differ in measurement andclassification; and in some instances this may affect results. It seemscommonsense that walking conducted locally may be affected byrespondents' perceptions of the neighborhood, whereas walkingconducted outside the local area may not (Shenassa et al., 2006).Despite some exceptions (Humpel et al., 2004; Pikora et al., 2006;Suminski et al., 2005), most research does not explicitly state whetherthe respondent's walking took place in their neighborhood. Moreover,the majority of the research recorded PA (usually inclusive of walking).Again, it is conceivable that only PA conducted in neighborhood publicspaces would be limited by perceptions of crime and safety. Gymworkouts, attending dance classes or gardening at homemay contributetowards a respondent's PA classification, but these activities are unlikelyto be seriously constrained by neighborhood crime.

Conclusion

To date there is insufficient evidence to conclude that crime-related safety influences PA. Nonetheless, the results presentedsuggest that, particularly for women and older adults, crime-relatedsafety may constrain PA.

This review highlights some limitations and inconsistencies withthe research to date. In particular, the conceptualization andmeasurement of safety requires attention. Much of the research relieson inadequate measures of crime, which do not make explicit thesource of insecurity; and judgments about crime may not influencebehavior unless the perceived crime also causes an emotionalreaction. Moreover, real or perceived neighborhood safety may onlyconstrain PA that takes place in local public spaces, drawing intoquestion the relevance of some PA outcomes.

Many neighborhood characteristics may not directly affect PA;however these features may influence crime and residents' percep-tions of safety. Future research might explore direct and indirecteffects on PA, by exploring how social and built environmentcharacteristics influence residents' safety, and how safety mediatesthe relationship between environmental factors and PA. Finally, allstudies examining safety and PA in this review employed cross-sectional study designs. Longitudinal studies would assist in drawingconclusions about causality. Consideration of these recommendationscould enhance the evidence base and provide insights into appro-priate interventions.

Acknowledgments

This research was funded by an Australian Research Council Grant(#LP0455453). The first author is supported by an Australian ResearchCouncil Postgraduate Award (Industry), with the support of theDepartment for Planning and Infrastructure (DPI) as Industry Partner.The second author is supported by a NHMRC Research Fellowship(#503712). There are no potential conflicts of interest known to any ofthe contributing authors.


References

Ainsworth, B., Wilcox, S., Thompson, W., Richter, D., Henderson, K., 2003. Personal,social, and physical environmental correlates of physical activity in African-American women in South Carolina. Am. J. Prev. Med. 25, 23–29.

Angel, S., 1968. Discouraging Crime Through City Planning. Berkeley, Institute of Urbanand Regional Development.

Armstrong, T., Bauman, A., Davies, J., 2000. Physical activity patterns of australianadults. Results of the 1999 National Physical Activity Survey. National Institute ofHealth and Welfare.

Austin, D.M., Furr, L.A., Spine, M., 2002. The effects of neighborhood conditions onperceptions of safety. J. Crim. Justice 30, 417–427.

Ball, K., Timperio, A., Salmon, J., Giles-Corti, B., Roberts, R., Crawfords, D., 2007. Personal,social and environmental determinants of educational inequalities in walking: amultilevel study. J. Epidemiol. Comm. Health 61, 108–114.

Booth, M.L., Owen, N., Bauman, A., Clavisi, O., Leslie, E., 2000. Social–cognitive andperceived environment influences associated with physical activity in olderAustralians. Prev. Med. 31, 15–22.

Brown, B.B., Perkins, D.D., Brown, G., 2004. Crime, new housing, and housing incivilities in afirst-ring suburb:multilevel relationshipsacross time.Hous. PolicyDebate15, 301–345.

Brownson, R.C., Baker, E.A., Housemann, R.A., Brennan, L.K., Bacak, S.J., 2001.Environmental and policy determinants of physical activity in the United States.Am. J. Public Health 91, 1995–2003.

Carnegie, M.A., Bauman, A., Marshall, A., Mohsin, M., Westley-Wise, V., Booth, M.L.,2002. Perceptions of the physical environment, stage of change for physical activity,and walking among Australian adults. R. Q. Exerc. Sport 73, 146–155.

Centre for Disease Control and Prevention, 2001. Physical activity trends —United States, 1990–1998. Morb. Mort. Wkly. Rep. 50, 166–169.

Corti, B., Donovan, R., Holman, C., 1996. Factors influencing the use of physical activityfacilities: results from qualitative research. Health Promot. J. Aust. 6, 16–21.

Covington, J.M., Taylor, R.B., 1991. Fear of crime in urban residential neighborhoods:implications of between and within-neighborhood sources for current models.Sociol. Q. 32, 231.

Craig, C.L., Brownson, R.C., Cragg, S.E., Dunn, A.L., 2002. Exploring the effect of theenvironment on physical activity: a study examining walking to work. Am. J. Prev.Med. 23, 36–43.

De Bourdeaudhuij, I., Sallis, J.F., Saelens, B.E., 2003. Environmental correlates of physicalactivity in a sample of Belgian adults. Am. J. Health Promot. 18, 83–92.

DeFrances, C.J., Titus, R.M., 1993. Urban planning and residential burglary outcomes.Landsc. Urban Plan. 26, 179–191.

Doyle, S., Kelly-Schwartz, A., Schlossberg, M., Stockard, J., 2006. Active communityenvironments and health: the relationship of walkable and safe communities toindividual health. J. Am. Plan. Assoc. 72, 19–31.

Duncan, M., Mummery, K., 2005. Psychosocial and environmental factors associated withphysical activity among city dwellers in regional Queensland. Prev. Med. 40, 363–372.

Ellaway, A., Macintyre, S., Bonnefoy, X., 2005. Graffiti, greenery and obesity in adults:secondary analysis on European cross sectional survey. BMJ 331 (7517), 611–612.

Evenson, L.K., Sarmiento, O.L., Tawney, K.W., Macon, M.L., Ammerman, A.S., 2003.Personal, social, and environmental correlates of physical activity in North CarolinaLatina immigrants. Am. J. Prev. Med. 25, 77–85.

Eyler, A.A., Baker, E.A., Cromer, L., King, A.C., Brownson, R.C., Donatelle, R.J., 1998.Physical activity and minority women: a qualitative study. Health Educ. Behav. 25,640–652.

Eyler, A.A., Brownson, R.C., Bacak, S.J., Houseman, R.A., 2003. The epidemiology ofwalking for physical activity in the United States. Med. Sci. Sports Exerc. 1529–1536.

Ferraro, K.F., 1995. Fear of Crime: Interpreting Victimization Risk. New York, StateUniversity of New York Press.

Ferraro, K.F., LaGrange, R.L., 1987. The measurement of fear of crime. Sociol. Inq. 57,70–101.

Foster, C., Hillsdon, M., Thorogood, M., 2004. Environmental perceptions and walking inEnglish adults. J. Epidemiol. Commu. Health 58, 924–928.

Giles-Corti, B., Donovan, R.J., 2002a. The relative influence of individual, social and physicalenvironment determinants of physical activity. Soc. Sci. Med. 54, 1793–1812.

Giles-Corti, B., Donovan, R.J., 2002b. Socioeconomic status differences in recreationalphysical activity levels and real and perceived access to a supportive physicalenvironment. Prev. Med. 35, 601–611.

Hale, C., 1996. Fear of crime: a review of the literature. Int. Rev. Vict. 4, 79–150.Harrison, R.A., Gemmell, I., Heller, R.F., 2007. The population effect of crime and

neighbourhood on physical activity: an analysis of 15 461 adults. J. Epidemiol.Comm. Health 61, 34–39.

Herbert, D., Davidson, N., 1994. Modifying the built environment: the impact ofimproved street lighting. Geoforum 25, 339–350.

Hoehner, C.M., Brennan Ramirez, L.K., Elliott, M.B., Handy, S.L., Brownson, R.C., 2005.Perceived and objective environmental measures and physical activity amongurban adults. Am. J. Prev. Med. 28, 105–116.

Hooker, S.P., Wilson, D.K., Griffin, S.F., Ainsworth, B.E., 2005. Perceptions of environ-mental supports for physical activity in African American and White adults in arural county in South Carolina. Prev. Chronic Dis. 2, 1–10.

Humpel, N., Owen, N., Iverson, D., Leslie, E., Bauman, A., 2004. Perceived environmentattributes, residential location, and walking for particular purposes. Am. J. Prev.Med. 26, 119–125.

Humpel, N., Owen, N., Leslie, E., 2002. Environmental factors associated with adults'participation in physical activity: a review. Am. J. Prev. Med. 22, 188–199.

Huston, S., Evenson, K., Bors, P., Gizlice, Z., 2003. Neighborhood environment, access toplaces for activity, and leisure-time physical activity in a diverse North Carolinapopulation. Am. J. Health Promot. 18, 58–69.

Jacobs, J., 1962. The Death and Life of Great American Cities. Jonathon Cape, London.Kawachi, I., Kennedy, B.P., Wilkinson, R.G., 1999. Crime: social disorganization and

relative deprivation. Soc. Sci. Med. 48, 719–731.King, A.C., Castro, C., Wilcox, S., Eyler, A.A., Sallis, J.F., Brownson, R.C., 2000. Personal and

environmental factors associated with physical inactivity among different racial–ethnic groups of USmiddle-aged and older-agedwomen. Health Psychol.19, 354–364.

Leslie, E., Saelens, B., Frank, L., Owen, N., Bauman, A., Coffee, N., Hugo, G., 2005.Residents' perceptions of walkability attributes in objectively different neighbour-hoods: a pilot study. Health Place 11, 227–236.

Lewis, D.A., Maxfield, M.G., 1980. Fear in the neighborhoods: an investigation of theimpact of crime. J. Res. Crime Delinq. 160–189.

Li, F., Fisher, K.J., Brownson, R.C., Bosworth, M., 2005. Multilevel modelling of builtenvironment characteristics related to neighbourhood walking activity in olderadults. J. Epidemiol. Comm. Health 59, 558–564.

Lim, K., Taylor, L., 2005. Factors associated with physical activity among older people—apopulation-based study. Prev. Med. 40, 33–40.

Mayhew, P., Clarke, R., Burrows, J., Hough, J., Winchester, S., 1979. Crime in Public View:Home Office Research Study No. 49. Her Majesty's Stationary Office, London.

McCormack, G., Giles-Corti, B., Lange, A., Smith, T.,Martin, K., Pikora, T.J., 2004. Anupdate ofrecent evidence of the relationship between objective and self-report measures of thephysical environment and physical activity behaviours. J. Sci. Med. Sport 7, 81–92.

Miles, R., Panton, L., 2006. The influence of the perceived quality of communityenvironments on low-income womens efforts to walk more. J. Comm. Health 1–14.

Mota, J., Lacerda, A., Santos, M.P., Ribeiro, J.C., Carvalho, J., 2007. Perceived neighborhoodenvironments and physical activity in an elderly sample. Percept. Mot. skills 104,438–444.

Newman, O., 1972. Defensible Space— People and Design in the Violent City. Macmillan,New York.

Owen, N., Humpel, N., Leslie, E., Bauman, A., Sallis, J.F., 2004. Understanding environmentalinfluences onwalking: review and research agenda. Am. J. Prev. Med. 27, 67–76.

Painter, K., 1996. The influence of street lighting improvements on crime, fear andpedestrian street use, after dark. Landsc. Urban Plan. 35, 193–201.

Parks, S.E., Housemann, R.A., Brownson, R.C., 2003. Differential correlates of physicalactivity in urban and rural adults of various socioeconomic backgrounds in theUnited States. J. Epidemiol. Comm. Health 57, 29–35.

Pikora, T.J., Giles-Corti, B., Knuiman, M.W., Bull, F., Jamrozik, K., Donovan, R.J., 2006.Neighborhood environmental factors correlated with walking near home: usingSPACES. Med. Sci. Sports Exerc. 38, 708–714.

Piro, F.N., Noess, O., Claussen, B., 2006. Physical activity among elderly people in a citypopulation: the influence of neighbourhood level violence and self perceived safety.J. Epidemiol. Comm. Health 60, 626–632.

Riger, S., LeBailly, R.T., Gordon, M.T., 1981. Community ties and urbanites' fear of crime:an ecological investigation. Am. J. Comm. Psychol. 9, 653–665.

Ross, C.E., 1993. Fear of victimization and health. J. Q. Criminol. 9, 159–175.Ross, C.E., 2000. Walking, exercising and smoking: does neighborhood matter? Soc. Sci.

Med. 51, 265–274.Sallis, J.F., Johnson,M.F., Calfas, K.J., Caparosa, S., Nichols, J.F.,1997. Assessing perceived physical

environmental variables thatmay influence physical activity. Res. Q. Exerc. Sport 68, 345.Sallis, J.F., Bauman, A., Pratt, M., 1998. Environmental and policy interventions to

promote physical activity. Am. J. Prev. Med. 15, 379–397.Sampson, R.J., Raudenbush, S.W., Earls, F., 1997. Neighborhoods and violent crime: a

multilevel study of collective efficacy. Science 277, 918–924.Shenassa, E.D., Leibhaber, A., Ezeamama, A., 2006. Perceived safety of area of residence

and exercise: a pan-European study. Am. J. Epidemiol. 163, 1012–1017.Skogan, W.G., 1990. Disorder and Decline: Crime and the Spiral of Decay in American

Neighborhoods. The Free Press, New York.Skogan, W.G., Maxfield, M.G., 1981. Coping with Crime: Individual and Neighborhood

Reactions. Sage Publications, Beverley Hills.Sooman, A., Macintyre, S., 1995. Health and perceptions of the local environment in

socially contrasting neighbourhoods in Glasgow. Health Place 1, 15–26.Steptoe, A., Feldman, P., 2001. Neighborhood problems as sources of chronic stress:

development of a measure of neighborhood problems, and associations withsocioeconomic status and health. Ann. Behav. Med. 23, 177–185.

Stokols,D.,1992. Establishingandmaintaininghealth environments. Am. Psychol. 47, 6–22.Suminski, R.R., Walker, S., Poston, C., Petosa, R., Stevens, E., Katzenmoyer, L., 2005.

Features of the neighborhood environment and walking by US adults. Am. J. Prev.Med. 28, 149–155.

Taylor, R.B., Koons, B., Kurtz, E., Greene, J., Perkins, D.D., 1995. Street blocks with morenonresidential land use havemore physical deterioration. Urban Aff. Rev. 31,120–136.

Troped, P.J., Saunders, R.P., Pate, R.R., Reininger, B., Addy, C.L., 2003. Correlates ofrecreational and transportation physical activity among adults in a New Englandcommunity. Prev. Med. 37, 304–310.

U.S. Department of Health and Human Services, 1996. Physical activity and health: areport of the surgeon general. Atlanta, GA., U.S. Department of Health and HumanServices: Centres for Disease Control and Prevention.

van Lenthe, F.J., Brug, J., Mackenbach, J.P., 2005. Neighbourhood inequalities in physicalinactivity: the role of neighbourhood attractiveness, proximity to local facilities andsafety in the Netherlands. Soc. Sci. Med. 60, 763–775.

Vest, J., Valadez, A., 2005. Perceptions of neighbourhood characteristics and leisure-time physical inactivity — Austin/Travis County, Texas, 2004. Morb. Mort. Wkly.Reports 54, 926–928.

Voorhees, C.C., Young, D.R., 2003. Personal, social and physical environmental correlatesof physical activity levels in urban Latinas. Am. J. Prev. Med. 25, 61–68.

Vrij, A., Winkel, F.W., 1991. Characteristics of the built environment and fear of crime: aresearch note on interventions in unsafe locations. Dev. Behav. An InterdisciplinaryJournal 12, 203–215.


Weinstein, A., Feigley, P., Pullen, P., Mann, L., Redman, L., 1999. Neighborhood safety andthe prevalence of physical inactivity— selected states. J. Am. Med. Assoc. 281, 1373.

Wilbur, J., Chandler, P.J., Dancy, B., Lee, H., 2003a. Correlates of physical activity in urbanmidwestern African-American women. Am. J. Prev. Med. 25, 45–52.

Wilbur, J., Chandler, P.J., Dancy, B., Lee, H., 2003b. Correlates of physical activity in urbanmidwestern Latinas. Am. J. Prev. Med. 25, 69–76.

Wilcox, S., Castro, C., King, A.C., Housemann, R., Brownson, R., 2000. Determinants ofleisure time physical activity in rural compared with urban older and ethnicallydiverse women in the United States. J. Epidemiol. Comm. Health 54, 667–672.

Wilcox, S., Bopp, M., Oberrecht, L., Kammermann, S., McElmurray, C., 2003. Psychosocialand perceived environmental correlates of physical activity in rural and olderAfrican-American and White women. J. Gerontol. Psychol. Sci. 58, 329–337.

Wilson, D.K., Kirkland, K.A., Ainsworth, B.E., Addy, C.L., 2004. Socioeconomic statusand perceptions of access and safety for physical activity. Ann. Behav. Med. 28,20–28.

Wood, L., Shannon, T., Bulsara, M., Pikora, T., McCormack, G., Giles-Corti, B., 2008. Theanatomy of the safe and social suburb: an exploratory study of the builtenvironment, social capital and residents' perceptions of safety. Health Place 14,15–31.

World Health Organization, 2002. The world health report— reducing risks, promotinghealthy life. World Health Organization, Geneva.

Young, D.R., Voorhees, C.C., 2003. Personal, social, and environmental correlates ofphysical activity in urban African-American women. Am. J. Prev. Med. 25, 38–44.

Zelinka, A., Brennan, D., 2001. Safescape. American Planning Association, Chicago.

the built environment, neighborhood crime and constrained physical activity: an exploration of...

Documents