lable at ScienceDirect

Journal of Environmental Psychology 31 (2011) 79e88

Contents lists avai

Journal of Environmental Psychology

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

Creating safe walkable streetscapes: Does house design and upkeep discourageincivilities in suburban neighbourhoods?

Sarah Foster*, Billie Giles-Corti, Matthew KnuimanCentre for the Built Environment and Health, School of Population Health, The University of Western Australia, 35 Stirling Highway, Crawley, WA 6009, Australia

a r t i c l e i n f o

Article history:Available online 2 April 2010

* Corresponding author. Tel.: þ61 8 6488 8730; faxE-mail addresses: sfoster@meddent.uwa.edu.au (

uwa.edu.au (B. Giles-Corti), matthew.knuiman@uwa.e

0272-4944/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.jenvp.2010.03.005

a b s t r a c t

There is growing evidence that residents are more likely to walk in attractive neighbourhoods, and thatnegative visual cues can deter residents from engaging in physical activity. This study explored thepremise that house design and upkeep could inhibit the incidence of physical disorder in suburbanstreets, thus contributing to a more pleasant walking environment for pedestrians. Street segments(n ¼ 443) in new residential developments (n ¼ 61) in Perth, Western Australia, were audited for houseattributes that facilitate natural surveillance (e.g., porch/verandah) or indicate territoriality (e.g., garden/lawn upkeep), and physical incivilities. A composite index of street-level house attributes yielded highlysignificant associations with disorder (trend test p ¼ 0.001) and graffiti (trend test p ¼ 0.005), signifyingthat the cumulative effect of several key attributes had greater potential to discourage incivilities in thestreet than any single characteristic. The findings suggest house design and upkeep may contribute to thecreation of safe, inviting streets for pedestrians.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Over the last decade, there has been increasing interest in thecharacteristics of the physical environment that encourage resi-dents to be physically active. Much of the research has focused onneighbourhood-level planning attributes (e.g., street connectivity,land-use mix, residential density) (Cerin, Leslie, du Toit, Owen, &Frank, 2007; Frank, Schmid, Sallis, Chapman, & Saelens, 2005;McCormack, Giles-Corti, & Bulsara, 2008; Owen et al., 2007), orthe presence of infrastructure that provides opportunities forphysical activity (e.g., footpaths, proximity of parks) (Giles-Corti &Donovan, 2002a; Owen, Humpel, Leslie, Bauman, & Sallis, 2004;Pikora et al., 2006). However, evidence also suggests that resi-dents are more likely to walk in attractive neighbourhoods, whereyards are well kept and the streetscape and buildings are diverseand interesting (Ball, Bauman, Leslie, & Owen, 2001; Giles-Corti &Donovan, 2003; Giles-Corti & Donovan, 2002a, 2002b; Michael,Green, & Farquhar, 2006; Owen et al., 2004; Pikora et al., 2006;Sugiyama, Leslie, Giles-Corti, & Owen, 2009). Furthermore, thereis support for the notion that negative visual cues, which detractfrom the neighbourhood’s presentation, can deter residents from

: þ61 8 6488 1199.S. Foster), billie.giles-corti@du.au (M. Knuiman).

All rights reserved.

engaging in physical activity (Ellaway, Macintyre, & Bonnefoy,2005; King, 2008; Mendes de Leon et al., 2009; Miles, 2008;Nagel, Carlson, Bosworth, & Michael, 2008; Shenassa, Leibhaber,& Ezeamama, 2006; Sugiyama & Ward Thompson, 2008).

The influence of positive and negative neighbourhood qualitieson physical activity is exemplified in a study by Ellaway et al.(2005). European residents in neighbourhoods with high levels ofobjectively recorded disorder (i.e., litter, graffiti, and dog mess) had50% lower odds of being physically active, whereas those in areaswith more vegetation and greenery had three times higher odds ofbeing active (Ellaway et al., 2005). In another example, Borst,Miedema, de Vries, Graham, and van Dongen (2008) invited olderadults to record the location and attractiveness of their walkingroutes, and then assessed their characteristics. Attractiveness ten-ded to be associated with pedestrian infrastructure, aestheticqualities and activity generating land-uses (ensuring the presenceof people). Litter detracted from attractiveness, whereas frontgardens enhanced attractiveness (Borst et al., 2008).

Neighbourhood incivilities (e.g., litter, graffiti and vandalism)may directly affect residents walking behaviour by detracting fromthe aesthetic appeal of public space e making walking lesspleasant. Moreover, the capacity for incivilities to inhibit physicalactivity and walking may be mediated by perceptions of crime andsafety (Foster & Giles-Corti, 2008; Loukaitou-Sideris & Eck, 2007;Ross, 1993). Incivilities reflect a breakdown of social control andhave been associated with higher crime (Brown, Perkins, & Brown,

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e8880

2004b), perceived crime (Perkins, Meeks, & Taylor,1992) and fear ofcrime (Robinson, Lawton, Taylor, & Perkins, 2003). If incivilities arenot promptly repaired, the ‘broken windows’ hypothesis suggeststhat disorder will escalate, initiating a spiral of fear and communitymistrust, which weakens social control, allowing incivilities andcrime to proliferate (Skogan, 1990). In addition, house and gardenmaintenance have been conceptualised as ‘suburban incivilities’and have been associated with crime and perceived safety (Brown,Perkins, & Brown, 2004a).

There are potential synergies between the built environmentattributes that contribute to an aesthetically pleasing streetscapefor walkers and those that deter disorder and crime. For instance,Craig, Brownson, Cragg, and Dunn (2002) rated streets for variousthemes including: (1) safety from crime; and (2) potential forcrime. Both contributed to a summary environmental score thatwas associated with walking to work. The crime-related elementsincluded lighting, front porches, escape routes, property mainte-nance, graffiti, vandalism, disrepair, and the presence of otherpeople (Craig et al., 2002). These themes are intrinsic to CrimePrevention through Environmental Design (CPTED) (Crowe, 1991),which seeks to deter offenders by creating a built environment thatincreases opportunities for surveillance, distinguishes public fromprivate territory, and presents a ‘positive image’. In combination,the CPTED principles aim to produce a landscape where offendersare at greater risk of being seen (i.e., natural surveillance), andspace is well-maintained, signifying a proprietary attitude over theenvironment (i.e., territoriality) (Brown et al., 2004a; Cozens,Saville, & Hillier, 2005). CPTED principles are increasingly beingembedded into planning and design guidelines for new andestablished urban development (Cozens et al., 2005; Raco, 2007).However, despite this focus on physical design, the efficacy ofenvironmental measures as a means to limit crime and alleviatefeelings of insecurity is intertwined with the social environment(Cozens et al., 2005; Merry, 1981). The reciprocity between thephysical and social environments sees residents both shape theirneighbourhoods and respond to cues in the physical environment(Stokols, 1992; Wood & Giles-Corti, 2008).

Evidence supporting the notion that environmental design candiscourage incivilities and affect the quality of public space islimited, as most studies examine crime or fear of crime. Nonethe-less, disorder and crime form part of the same spectrum, andalthough most disorder offences are victimless and relativelyminor, they threaten that there is ‘potential for harm’ (Ross,Mirowsky, & Pribesh, 2002, p. 63). The most prominent studyexploring building design and disorder is Coleman’s (1985) study ofresidential buildings and social malaise.While primarily focused onhigher density council towers, single family housing was also ratedfor physical disorder. Although methodological concerns have beenraised (Campbell, 1985; Hillier, 1986), Coleman’s findings revealthat houses with well-defined private territory and good potentialfor street surveillance were associated with less litter (Coleman,1985). Other research highlights the relevance of building designelements (e.g., personalised decoration, street lighting andsurveillance opportunities) to residents’ perceptions of incivilities(Perkins et al., 1992). These studies hint that house characteristicshave some potential to discourage physical disorder.

More commonly, studies have examined characteristics of thebuilt environment for associations with crime. In an urban context,houses with porches and a nearby store were positively associatedwith crime. This finding was somewhat counterintuitive, sinceporches theoretically promote natural surveillance; however thestudy was set in undesirable city neighbourhoods comprisingsomewhat transient populations (Schweitzer, Kim, & Macklin,1999). Indeed, it has been suggested that the capacity for envi-ronmental design to prevent crime may be more effective in stable

neighbourhoods with higher community integration (Booth, 1981).Studies in suburban settings suggest house attributes do play somerole in restricting crime. Brown and Altman (1983) found physicalbarriers (e.g., fences and locked gates); symbolic barriers (e.g.,evidence the homewas occupied, personalised decoration) and thepotential for surveillance (e.g., visibility from neighbouring prop-erties) were protective against home burglary. Moreover, a studylocated in a declining suburban neighbourhood found ‘home inci-vilities’ (an objective measure encompassing incivilities, housingcondition and garden presentation) were independently associatedwith crime (Brown et al., 2004b).

Further studies suggest that house and street maintenancemight promote feelings of security. Wood et al. (2008) linked a highlevel of neighbourhood upkeep (e.g., maintained verges andgardens, minimal litter) with greater feelings of safety and socialcapital among residents. Similarly, another study found objectivelyrated housing quality, as characterised by building maintenanceand upkeep, had a direct association with residents’ perceptions ofsafety, as well as an indirect effect via satisfaction with the physicaland social environment (Austin, Furr, & Spine, 2002). Furthermore,other neighbourhood cues have the potential to affect perceivedsafety. For instance, Schweitzer et al. (1999) found objectivelyrecorded neighbourhood watch signs correlated with fear of crime.Like physical incivilities, these signsmay signal to residents that thearea is unsafe.

Together, these findings suggest that house design and upkeepmay be an effective means of limiting disorder and crime, andimproving perceptions of safety in the suburban landscape. Theseattributes may also help create amore convivial street environmentand help promotewalking. However, research examining correlatesof disorder has largely focused on deteriorating urban centres, withsuburban environments receiving relatively sparse attention(Brown et al., 2004b; Phillips & Smith, 2006; Reisig & Cancino,2004). In particular, few empirical studies have tested the associ-ation between street-level house attributes and physical incivilitiesusing detailed objective data, and little is known about the cumu-lative effect of different house attributes. Previously cumulative riskmodels have been applied to simultaneously examine the envi-ronmental risk factors affecting health outcomes (Evans, 2003;Evans, Kim, Ting, Tesher, & Shannis, 2007; Evans & Marcynyszyn,2004). Similarly, individual house characteristics studied in isola-tion may not be associated with the incidence of disorder, whereasthe combined effect of several attributes may reveal a strongereffect.

Thus, this study aimed to examine current gaps in the literaturerelated to the impact of surveillance and territoriality in a suburbanlocale. We hypothesised that suburban streets comprised of houseswith more attributes that facilitate natural surveillance or reflectresidents’ territorial behaviour would be associated with lessevidence of physical incivilities in the street environment. More-over, we hypothesised there would be a doseeresponse relation-ship in the association between house attributes and incivilities.The study hypothesis and overall context are summarised in Fig. 1.

2. Methodology

2.1. Study context and data collection

The Neighbourhood Environment Safety Tool (NEST) wasdeveloped to collect objective audit data on the features ofsuburban streets that might influence incivilities, perceived safetyand walking among residents. The formatting and some items werederived from SPACES (Pikora et al., 2002) but ideas were alsosourced from other instruments (Barnes McGuire, 1997; Boarnet,Day, Alfonzo, Forsyth, & Oakes, 2006; Brownson et al., 2004;

Study hypothesis

Perceptions of

crime & safety

Incivilities in

suburban streets

Signs that

residents ‘care’

Potential for

surveillance

House design &

maintenance

Study context

Walking in the

neighbourhood

Fig. 1. The study hypothesis: that house design and maintenance are associated with the incidence of physical incivilities in suburban residential streets.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e88 81

Burton, 2005; Caughy, O’Campo, & Patterson, 2001; Dunstan et al.,2005; Lee, Booth, Reese-Smith, Regan, & Howard, 2005; Minnery,2005; Perkins et al., 1992; Raudenbush & Sampson, 1999;Winchester & Jackson, 1982; Zenk et al., 2007). In accordancewith previous research, NEST was designed to capture negativevisual cues and positive expressions of residents’ territorial func-tioning (Perkins et al.,1992).While there aremore novel techniquesto record audit data (Brownson et al., 2004; Raudenbush &Sampson, 1999; Zenk et al., 2007), the manual checklist is simpleto use, easy to manage in the field and requires limited resources.NEST is available on request from the first author.

The audited streets were drawn from new suburban residentialdevelopments, which had been established for between one andthree years, and were typically located on the urban fringe. Streetswere selected to correspond to the home addresses of a subset ofparticipants in the RESIDential Environments (RESIDE) project,a five-year longitudinal study evaluating the impact of a state-government sub-division code in Perth, Western Australia. Fulldetails are described elsewhere (Giles-Corti et al., 2008). The studywas approved by The University of Western Australia’s HumanResearch Ethics Committee.

The sampling strategy resulted in data specific to a subset ofparticipant’s immediate environment; however this paper focusesexclusively on the audit data. The focus on respondents’ residentialstreets was driven by the theoretical importance of the ‘homepatch’ to influence feelings of security (Barton, Grant, & Guise,2003), with events occurring within the block scale setting ofgreater significance to residents than those occurring in the widerneighbourhood (Taylor, 1987). Previous studies have used the streetblock (or block-face) as the basic unit for the analysis of surveil-lance, territorial markers and incivilities, on the grounds that theblock has a more distinct and easily defined boundary than theneighbourhood, and blocks tend to be culturally homogeneous andas such residents often share the same concerns as their neighbours(Barnes McGuire, 1997; Boarnet et al., 2006; Caughy et al., 2001;Franzini, Caughy, Nettles, & O’Campo, 2008; Perkins, Florin, Rich,Wandersman, & Chavis, 1990; Perkins et al., 1992; Pikora et al.,2002; Raudenbush & Sampson, 1999; Schweitzer et al., 1999).

During May and June 2007, between the hours of 9am and 3pmon weekdays, a team of three trained assessors audited 443 resi-dential segments (defined as the section of street between twointersections, including both sides of the street). Data werecollected during these times to avoid the higher traffic flows causedby travel to/from school and work, as items required auditors toassess traffic and pedestrian volumes. Segments were sampledfrom as many residential developments as possible (n ¼ 61) acrossdifferent suburbs (n ¼ 45) to minimise the influence of contextualneighbourhood-level factors on the associations between housecharacteristics and disorder. Between one and 34 segments(mean ¼ 7.26, standard deviation ¼ 5.79) were audited from eachhousing development, and each segment contained between twoand 48 houses (mean ¼ 12.27, standard deviation ¼ 7.25).

Auditors walked each segment to record data, rather thanauditing from a car window or video-taping, as this provideda better viewof the houses and street; andmany features of interest(e.g., litter, smaller graffiti or tags) were not easily visible from a car.Moreover, auditors on foot are viewed with less suspicion by resi-dents than those in vehicles (Caughy et al., 2001).

2.2. Predictor variables

Auditors counted the total number of residential lots in eachsegment, and recorded the design and maintenance attributes thatmight facilitate (or impede) natural surveillance, and indicate theterritorial functioning of householders. Attributes included: (1)good visibility from the street (i.e., windows clearly visible); (2)verandahs, porches or balconies; (3) double garages; (4) high, solidfront walls; (5) low front walls, fences, hedges or borders thatdefine private space without restricting visibility; (6) outdoorfurniture; (7) personalised decoration (e.g., elaborate letter boxes,house name plates); (8) security grills, bars or roller shutters onfront windows; (9) unkempt front lawns; (10) unkempt frontgardens; (11) vacant lots; and (12) houses under construction. Therationale for collecting these attributes is outlined below.

Natural surveillance in suburban neighbourhoods can bepromoted by house designs that encourage residents to observe thestreet while minimising obstacles that restrict visibility (Zelinka &Brennan, 2001). Front verandahs, porches or balconies increaseopportunities for residents’ to monitor the street and provide spacefor family and neighbour interaction (Brown, Burton, & Sweaney,1998). However, porch sizes have declined over time (or dis-appeared fromhouse designs altogether) (Brown et al., 1998), to thepoint where many new houses have a token porch, built for ‘kerbappeal’ but often too small to serve its traditional function(Kunstler, 1996). Consequently, only verandahs, porches or balco-nies where several people could comfortably sit together wererecorded. Other house features, such as garage doors and highwalls, have a dual role in limiting the potential for surveillance andpreventing resident interaction. The shrinking porch has beenoffset by larger, more prominent garages; and with increased caruse, the garage has migrated from the back corner to the housefront, becoming ‘a primary element of the streetscape’ (Southworth& Owens, 1993, p. 13). As a consequence, the presence of double,triple or even quadruple garages and high solid front walls cancreate a street façade dominated by blank wall (Brown et al., 1998).

Demonstrations of house personalisation, alterations andcondition can imply a connection to home and neighbourhood(Werner, Peterson-Lewis, & Brown, 1989), denote the strength ofa resident’s place attachment, and create a symbolic barrierbetween public space and private property (Brown et al., 2004b;Perkins et al., 1992). Brown and Altman (1983, p. 205) proposedthat the deterrent value of territorial markers ‘stems from thesymbolic communication of residents’ concern for and defence ofthe territory’. Again, the attractiveness of houses and yards has

Table 1Inter-rater reliability of study variables pertaining to street segments (n ¼ 69).

Characteristic Kappa

Independent variables (house characteristics)Houses with good visibility from the street(i.e., windows are clearly visible from street)

0.65

Houses with front verandah, porch or balcony 0.76Houses with double garages doors fronting the street 0.85Houses with a high solid front wall 0.97Houses with a low front wall, fence, hedge orborder marking the property

0.76

Houses with outdoor furniture in the front yard,verandah, porch or balcony

0.76

Houses with personalised decoration(e.g., garden ornaments, name plate)

0.65

Houses with security bars, grills or rollershutters on front windows

0.88

Houses with unkempt front lawns 0.43Houses with unkempt front gardens or no garden at all 0.62Vacant lots 0.94Houses under construction 0.74

Dependent variables (street characteristics)Disorder 0.65Graffiti 0.80

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e8882

a dual role as it can symbolise ‘accessibility to and cohesivenesswith neighbours’ (Werner et al., 1989, p. 285). For instance,neighbours construe residents with outdoor furniture on theirporch to be more sociable, and outsiders tend to interpretChristmas decorations as indicators of residents’ social integrationor intention to ‘interact’ with neighbours (Werner et al., 1989).

However, other elements of house presentation can communi-cate negative connotations. In a suburban context, well presentedgardens and lawns are typically a social norm, indicative ofcommitment (Jackson, 1978; Smardon, 1988); whereas unkemptyards can adversely affect the appearance of the street (Nassauer,Wang, & Dayrell, 2009), and reflect weaker neighbourhood bonds(Harris & Brown, 1996). Obvious security measures can also detractfrom the presentation of the house and streetscape, and conveya general mistrust of the local community (Ross et al., 2002). Takento excess, security precautions can produce a ‘fortress effect’, whichmay serve to fuel public perceptions of insecurity (Schneider &Kitchen, 2007) and counteract other household CPTED compo-nents (Cozens et al., 2005). Finally, vacant lots or houses underconstruction represent territorial gaps, where natural surveillanceis limited and there are no residents to actively manage the space.Such territorial gaps have been associated with a breakdown ofsocial control (Taylor, Koons, Kurtz, Greene, & Perkins, 1995) andhigher crime (Greenberg, Rohe, & Williams, 1982).

2.3. Outcome variables

The study outcomes were graffiti and disorder. These variablesare commonly merged into a ‘physical incivility’measure; howeverthe distinction was preserved in this study. In some ways graffiti isconceptually different (Austin & Sanders, 2007) and may be anoutlet for local teenagers in new suburbs with few recreationalfacilities (Kunstler, 1996), whereas other forms of disorder areperhaps less aligned with youth, and in some instances mayrepresent more aggressive (e.g., vandalism) or potentiallydangerous (e.g., burn outs/skid marks) behaviour.

Auditors recorded the presence of various forms of physicaldisorder. However, many items were not present (e.g., drug or sexparaphernalia, abandoned vehicles); and others were deemed notto detract from the aesthetic presentation of the street, or reflectanti-social behaviour. (e.g., cigarette butts, roaming dogs). The finaldisorder outcomewas a dichotomous variable, based onwhether atleast two of the following were recorded in the segment: (1) dis-carded alcohol containers; (2) broken glass; (3) vandalism (e.g.,damage to public infrastructure, front fences, letter boxes etc); and(4) vehicle skid marks. While it is plausible that some damage maybe the product of innocent events (e.g., vandalism or skid marksfrom car accidents) rather than destructive intent, these acts stillaffect the presentation of the public realm. Moreover, 98% of theaudited segments were exclusively residential (with restricted50 km per hour speed limits), suggesting that many of the recordedskid marks were a ‘visual legacy’ of intentional burn outs(Armstrong & Steinhardt, 2006; Falconer & Kingham, 2007). Graffitiwas recorded according to the number of sites in each segment andtheir size. For the purposes of this study, minor graffiti (i.e., paintedover graffiti; one or two small tags) were excluded; but multipletags and larger graffiti counted towards the dichotomous graffitioutcome. The treatment of both outcomes was intended to focusthe analyses on street segments with more prolific graffiti ordisorder.

2.4. Reliability testing

Following examples from previous research, (Pikora et al., 2002;Zenk et al., 2007) a series of measures were taken to ensure NEST

items had good inter-rater reliability. A comprehensive manual,including information on audit protocols, operational definitionsand photographs was developed to steer staff through the auditprocess and checklist. Each auditor attended an individual trainingsession, where theywere guided through the checklist andmanual.Auditors undertook practice audits during and after this session;their results were compared with the main researcher’s and anydiscrepancies discussed. Once in the field, auditors were encour-aged to telephone for additional guidance when uncertaintiesarose.

Features that remain stable over time generally have betterreliability than highly changeable features (e.g., dog litter),ambiguous features (e.g., personalised decoration) or those thatrequire some judgement on the part of the auditor (e.g., unkemptlawns or gardens) (Brownson et al., 2004; Zenk et al., 2007). Thisalso proved the case with NEST, which was tested for inter-raterreliability using a sample of 69 segments. Evidence of physicaldisorder was highly changeable (e.g., graffiti, alcohol containers,broken glass), and other elements could easily be overlooked (e.g.,dog litter, cigarette butts). Moreover, features that reflected thehouseholder’s territoriality were also prone to change and requiredsome judgement on the part of the auditor (e.g., unkempt gardensand lawns, personalised decoration). These items tended to havelower inter-rater reliability; however their stability improvedwhendata were collapsed into the summary variables which were usedin the analysis. As a result, most items proved to have substantialinter-rater reliability (Kappa 0.62e0.97); with the only exceptionbeing ‘unkempt front lawns’which hadmoderate reliability (Kappa0.43). However, it is not possible to identify whether any lack ofagreement was due to the different raters or changes in the envi-ronment over time. The final items used in the analysis and theirinter-rater reliability values are listed in Table 1.

2.5. Statistical analyses

Data were analysed using the Statistical Package for the SocialSciences (SPSS) version 15. Predictor variables were created bycalculating the proportion of houses in the street segment witheach feature and then dichotomising. House attributes witha normal distribution: (1) good visibility; (2) double garages; (3)low front walls, hedges or borders; (4) outdoor furniture; and (5)personalised decoration, were dichotomised at the median.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e88 83

However, some measures were heavily skewed, with mostsegments recording no houses with the relevant attribute (i.e.,median value ¼ 0). Hence, skewed variables were dichotomisedaccording to whether any houses in the segment had the feature.This applied to: (1) front verandahs, porches or balconies; (2) high,solid front walls; (3) security bars, grills or roller shutters on frontwindows; (4) unkempt front lawns, (5) unkempt gardens; (6)vacant lots; and (7) construction sites.

House attributes that held stronger theoretical and statisticalassociations with disorder and graffiti were combined into anadditive scale (negative items reversed). The ‘CPTED index’comprised six attributes: (1) houses with good visibility; (2)verandahs, porches or balconies; (3) low walls, fences, hedges orborders; (4) unkempt gardens (reversed); (5) unkempt lawns(reversed); and (6) the presence of vacant lots (reversed). The indexwas adopted to explore whether the odds of incivilities reducedwith each additional positive feature (or the absence of a negativefeature) in the segment. Visible security measures were notincluded in the index because of their potential to undermine theeffectiveness of other CPTED components (Cozens et al., 2005).

All analyses were conducted using logistic regression (general-ised estimating equations) with adjustment for clustering withinresidential development. Although all the variables of interest weremeasured at the individual (segment) level, adjustment for clus-tering was necessary because the more proximate street segments(co-located within residential developments) would most likelyhave similar characteristics (e.g., the presence of incivilities).Furthermore, all models controlled for the total number of resi-dential lots in each segment, because longer segments wouldincrease the probability of different house attributes being presentand increase the likelihood of incivilities. Additional adjustment forstreet lighting and assessments of pedestrian traffic volume madeno difference to the results. The main analyses were conducted infour stages:

1. The univariate associations between the house attributes andthe outcomes: (1) disorder; and (2) graffiti were examined;

2. These associations were adjusted for lot value, which was usedas a proxy measure for socio-economic status. Lot value wasderived from the Valuer General’s Office valuation of residen-tial lots, based on the property’s capital value at the time ofpurchase. For these analyses, lot values were aggregated acrosseach residential development using the full RESIDE studycompliment (mean valuation ¼ AUD$78,844), and then splitinto low, medium and high tertiles for analysis.

3. House attributes found to have a stronger relationship(p < 0.10) with disorder and graffiti were carried into themultivariate logistic regression. A backwards stepwise elimi-nation approach was used to reduce to a final model thatcontained variables significantly (p < 0.05) and independentlyassociated with the outcomes.

4. The CPTED index was included in logistic regression models toexplore whether house design and maintenance attributes hada cumulative affect on incivilities.

3. Results

3.1. Physical disorder

Approximately 28% of the 443 audited street segments hadphysical disorder. Table 2 shows that three attributes had signifi-cant univariate associations with disorder, after controlling for thenumber of residential lots and clustering within residential devel-opment (Model 1). Lot value was also associated with disorder;with lots on segments in higher priced developments having 65%

lower odds of physical disorder. When the house attributes werefurther adjusted for lot value (Model 2), and in the final multivar-iate analyses (Model 3), the previously observed associationsremained significant. There were significantly lower odds ofdisorder being present where at least one house on the segmenthad a verandah, porch or balcony (OR 0.60, 95% CI: 0.37e0.98); orthere was a higher proportion of houses with low front walls,fences, hedges or borders (OR 0.40, 95% CI: 0.25e0.65). Conversely,the likelihood of finding disorder significantly increased when thesegments had at least one vacant lot (OR 4.03, 95% CI: 2.35e6.92).

3.2. Graffiti

Almost 15% of the segments had graffiti present. Table 2demonstrates that two attributes had significant univariate asso-ciations with graffiti (Model 1). The odds of graffiti being present insegments was higher when at least one house had visible security(OR 1.70, 95% CI: 1.04e2.80), or a poorly maintained garden (OR2.36, 95% CI: 1.28e4.35). In a similar pattern to disorder, as lotvalues increased, the odds of graffiti being present fell. When thehouse attributes were further adjusted for lot value (Model 2),visible security attenuated to become non-significant, suggestingsome confounding. Separate analyses (not shown here) indicatedthat street segments from higher SES developments had signifi-cantly fewer visible security precautions. However, unkempt frontgardens remained significant after adjustment for lot values, andalso emerged as the only house attribute to be significantly asso-ciated with graffiti in the multivariate analysis (Model 3: OR 2.27,95% CI 1.21e4.28).

3.3. CPTED index

In the final analyses, the focus shifted to examine the cumulativeaffect of the six house attributes that formed the CPTED index. Fig. 2illustrates the association between the CPTED index and the odds ofphysical disorder and graffiti, after adjustment for the number ofresidential lots per segment, clustering within residential devel-opment and lot value. For each additional house attribute present(or the absence of a negative attribute), the odds of both physicaldisorder and graffiti decreased. Trend tests confirmed that the oddsof disorder reduced by approximately one third (OR 0.66, 95% CI0.54e0.80, p ¼ 0.000), and the odds graffiti fell by about onequarter (OR 0.74, 95% CI 0.61e0.92, p¼ 0.005), with each additionalCPTED characteristic.

4. Discussion

Althoughmany of the house attributes explored in this study areoften assumed to contribute to a reduction in crime and disorderand increase residents’ perceived safety, there is relatively sparseempirical research into this relationship in a suburban context. Thestudy findings confirmed the hypothesis that house attributes thatpromote natural surveillance and reflect residents’ territorialfunctioning are associated with a reduced odds of physical incivil-ities in suburban streets. Importantly from a policy and practiceperspective, the findings suggest that it is not simply the presenceof one or two attributes that have the greatest influence, but thecumulative effect of multiple features. The CPTED index revealedthat the odds of physical incivilities decreased with each additionalattribute. According to these analyses, houses in suburban streetswith the least physical incivility have good visibility, with at leastone house having a good sized verandah or balcony; a higherproportion of houses with low walls or borders; no unkemptgardens and lawns; and no vacant lots. Together, these house

Table 2The associations between house design and maintenance and the outcomes: (1) physical disorder and (2) graffiti for 443 audited street segments.

House characteristics % Physical disorder Graffiti

Model 1 Model 2 Model 3 Model 1 Model 2 Model 3

OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI)

Houses with good visibilityLess visibility 50.3 1.00 1.00 1.00 1.00More visibility 49.7 0.88 (0.58e1.33) 0.96 (0.63e1.46) 0.73 (0.46e1.16) 0.80 (0.51e1.28)

Porch, balcony or verandahNo verandah 55.3 1.00 1.00 1.00 1.00 1.001 þ verandah 44.7 0.51 (0.32e0.81)* 0.59 (0.37e0.92)* 0.60 (0.37e0.98)* 0.74 (0.42e1.30) 0.89 (0.50e1.61)

Double garage doors fronting streetLess 2� garages 47.9 1.00 1.00 1.00 1.00More 2� garages 52.1 0.75 (0.52e1.10) 0.86 (0.61e1.22) 0.73 (0.42e1.24) 0.87 (0.49e1.53)

High wallsNo high walls 31.8 1.00 1.00 1.00 1.001 þ high walls 68.2 1.23 (0.74e2.04) 1.18 (0.72e1.96) 1.21 (0.69e2.12) 1.14 (0.66e1.97)

Low wall, hedge or borderLess borders 47.6 1.00 1.00 1.00 1.00 1.00More borders 52.4 0.44 (0.27e0.71)* 0.46 (0.28e0.75)* 0.40 (0.25e0.65)* 0.67 (0.36e1.27) 0.72 (0.37e1.40)

Outdoor furnitureLess furniture 46.0 1.00 1.00 1.00 1.00More furniture 54.0 0.87 (0.62e1.22) 0.86 (0.61e1.21) 0.94 (0.63e1.41) 0.93 (0.62e1.39)

Personalised decorationLess decoration 49.0 1.00 1.00 1.00 1.00More decoration 51.0 0.75 (0.53e1.06) 0.78 (0.56e1.09) 0.94 (0.61e1.45) 1.00 (0.66e1.52)

Security bars/grills on front windowsNo visible security 63.7 1.00 1.00 1.00 1.001 þ visible security 36.3 1.29 (0.82e2.02) 1.14 (0.71e1.85) 1.70 (1.04e2.80)* 1.54 (0.92e2.56)

Unkempt front lawnsNo unkempt lawns 67.3 1.00 1.00 1.00 1.001 þ unkempt lawns 32.7 1.22 (0.82e1.81) 1.25 (0.84e1.86) 1.41 (0.87e2.28) 1.45 (0.87e2.42)

Unkempt front gardensNo unkempt gardens 57.1 1.00 1.00 1.00 1.00 1.001 þ unkempt gardens 42.9 1.52 (0.93e2.46) 1.42 (0.87e2.30) 2.36 (1.28e4.35)* 2.27 (1.21e4.28)* 2.27 (1.21e4.28)*

Vacant lotsNo vacant lots 55.3 1.00 1.00 1.00 1.00 1.001 þ vacant lots 44.7 2.90 (1.80e4.69)* 3.34 (2.02e5.53)* 4.03 (2.35e6.92)* 1.34 (0.85e2.12) 1.50 (0.86e2.34)

Houses under constructionNo houses 65.2 1.00 1.00 1.001 þ houses 34.8 1.20 (0.79e1.82) 1.31 (0.86e1.98) 1.06 (0.65e1.74) 1.19 (0.71e1.99)

Lot valueLowest tertile 30.7 1.00 1.00 1.00 1.00Medium tertile 36.8 0.51 (0.26e1.00)* 0.55 (0.29e1.04) 0.44 (0.20e0.97)* 0.46 (0.21e0.99)*Highest tertile 32.5 0.35 (0.18e0.66)* 0.32 (0.17e0.60)* 0.22 (0.09e0.51)* 0.24 (0.10e0.55)*

Model 1: Univariate associations: adjusted for number of residential lots & clustering by residential development. Model 2: Univariate associations: adjusted for numberresidential lots, clustering by residential development & lot value. Model 3: Multivariate model: adjusted for number residential lots, clustering by residential development &lot value.*p value <0.05.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e8884

design and upkeep attributes appear to help discourage expres-sions of disorder.

Although many attributes had some correlationwith incivilities,few individual attributes were independently associated. Someattributes emerged as important in accordance with the literature(e.g., verandah, porch or balcony, vacant lots), however otherswarrant discussion, particularly the strong association betweensegments with lowwalls or borders and reduced odds of disorder. Itis reasonable that better distinguished private space may detertrespassers and help restrict home burglary (Brown & Altman,1983), but how can the clear distinction of private space helpinhibit physical disorder in the public realm? One possible expla-nation is that the presence of some barrier designating the frontproperty line creates a street with a more intimate human scale,even in a suburban residential context. Without this border

distinguishing space, private territory is effectively marked by thehouse itself, which results in a much broader, possibly morealienating streetscape (i.e., the pedestrian feels distant from privateproperty). This is particularly true of neighbourhoods with widestreets and large lots, which ‘contribute to a less inviting pedestrianenvironment’ (Southworth & Owens, 1993, p. 282). When a highproportion of houses clearly mark the front property boundary,pedestrians walking along the footpath have a constant physicalreminder of the immediacy of private space. This visual cue mayhave some deterrent effect. For example, people may be lessinclined to litter when the street configuration makes them feelcloser to private property than distant from it.

Previous research has linked garden upkeep with fewerperceived incivilities (Taylor, 1988) and greater feelings of safety(Wood et al., 2008). This study highlighted an independent

0

0.2

0.4

0.6

0.8

1

1.2

0 or 1 2 3 4 5 6CPTED index

Od

ds

Disorder

Graffiti

Fig. 2. The CPTED index and odds of physical disorder and graffiti from fitted logisticmodels.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e88 85

association between houses with unkempt front gardens (or nogarden) and graffiti. As discussed earlier, potential offenders aretheorised to interpret these territorial displays as signals of placeattachment, a proprietary attitude and even the resident’s vigi-lance, thus deterring property offences (Brown & Altman, 1983).This rationale may also extend to disorder offences in the streetenvironment (Taylor, 1988). It is plausible that outsiders respond tostronger display of territoriality (and the inference that residentswould react to their presence) by curbing inappropriate behaviour.Standards of maintenance may provide visual cues as to whatbehaviour will be tolerated in the adjacent public spaces. However,it is equally possible that the observed association is a function ofimproved street surveillance, as residents with well-maintainedgardens would spend more time in their front yards (Brunson, Kuo,& Sullivan, 2001; Kuo & Sullivan, 2001).

Visible security was associated with greater odds of graffiti, butattenuated after adjustment for lot value. Separate analysisconfirmed that segments from low to middle SES residentialdevelopments were more likely to have obvious security. Theseprecautions are most likely installed in response to perceived crimeand incivilities (Caughy et al., 2001; Perkins et al., 1992). Howeverresidents that offset concerns about crime by fortifying their homescould paradoxically diminish perceptions of safety by underminingthe positive attributes of the streetscape (Cozens et al., 2005), sig-nalling their suspicions to the community (Doeksen, 1997) andintroducing another visual cue to suggest the area is unsafe (Foster& Giles-Corti, 2008). Indeed, there is some evidence supporting thenotion that visible security inflames residents fears. King (2008)found that, for a small sample of older adults, window bars werenegatively associated with total physical activity; but this associa-tion was largely mediated by perceived safety.

Consistent with other studies, the findings support the estab-lished connection between neighbourhood SES and physical inci-vilities (Caughy et al., 2001; King, 2008; Lee et al., 2005). Thispattern is noteworthy, given our streets are drawn from newsuburban developments, predominantly comprised of owner-occupiers, and therefore do not encompass areas of economicdisadvantage. Residents in new housing may have higher placeattachment which ‘may guard against incivilities as residentsremove litter, trim lawns and otherwise keep up appearances ofplaces that are sources of pride and identity’ (Brown et al., 2004b, p.361). However, as the association between house attributes andincivilities was independent of SES, and lower SES neighbourhoodstypically have higher levels of physical disorder (Caughy et al.,2001; Lee et al., 2005), it is conceivable that disadvantagedneighbourhoods might gain most from increased attention tohouse design and maintenance. Future research might examine

whether the study findings hold for lower SES neighbourhoods. Ifso, they pose a potential strategy to improve the quality of publicspace. An economic evaluation of CPTED house attributes meritsconsideration to assess whether the overall benefits to thecommunity outweigh the implementation costs.

It has long been recognised that the built environment alonewill not create safer communities (Booth, 1981; Merry,1981; Taylor,Gottfredson, & Brower, 1980). This study focused exclusively onphysical environmental characteristics and objective indicators ofincivility. However an alternative pathway must be acknowledged.Many of the house attributes conceptualised as promotingsurveillance or reflecting propriety may also enhance the potentialfor resident interaction, and even symbolise a willingness tosocialise with neighbours (Brown et al., 2009; Werner et al., 1989).For example, several house attributes have a twin function: a porchmay encourage surveillance, but also promotes interaction withneighbours (Brown et al., 1998); personalised house decorationssymbolise territoriality, but also infer residential attachment andfriendliness to neighbours (Werner et al., 1989); lawn and gardenupkeep help define private space but also reflect social cohesionthrough shared standards of home maintenance (Greenberg et al.,1982; Nassauer et al., 2009). Indeed, since physical incivilitiesindicate a breakdown of social control (Skogan, 1990) andcommunity mistrust (Ross et al., 2002), neighbourhood socialcohesion may mediate the association between the house attri-butes and physical incivilities.

The broader context for this study was to explore features of thestreetscape that could affect the pedestrian experience, withparticular regard to perceptions of crime and safety. Neighbour-hood streets are the most commonly used setting for engaging inphysical activity (Giles-Corti & Donovan, 2002a; Sugiyama et al.,2009), however the street characteristics that encourage recrea-tional activity are not well understood (Sugiyama et al., 2009). Ifattributes of the built environment can help kerb physical incivil-ities, improving the aesthetic quality of the public realm, they arealso worthy of further investigation as correlates of walking. Housedesign and maintenance could potentially influence walking viatwo pathways: (1) by improving the aesthetic presentation of thestreet, making walking more pleasant; or (2) by influencingperceptions of safety either directly (through potential forsurveillance and maintenance standards) or indirectly (throughtheir capacity to discourage expressions of antisocial behaviour).Indeed, there is evidence to support both pathways. One studyhighlighted a direct association between incivilities and physicalactivity (which was not mediated by perceived safety) (Miles,2008), whereas another identified perceived safety as a mediatorthe association between objective physical incivilities and totalphysical activity (King, 2008).

4.1. Limitations

This study has several limitations. Many house characteristicswere associated with the presence of incivilities; however fewindividual attributes achieved statistical significance. A largersample size might have improved the significance of these associ-ations, particularly for graffiti, which was identified in a smallersample of streets. Furthermore, graffiti and disorder were analysedas separate outcomes. The findings emphasised that more indi-vidual house attributes were associated with disorder than graffiti,and highlighted a slightly stronger association between the CPTEDindex and disorder. While both graffiti and disorder are con-ceptualised as physical incivilities, they perhaps reflect slightlydifferent elements of anti-social behaviour. Graffiti may simplyindicate bored teenagers, and graffiti writers may achieve somepersonal gratification from exposing their graffiti to as many people

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e8886

as possible. As a result, indicators of territoriality and the risk ofobservation may be less of a deterrent. Paradoxically, graffiti mayindicate a degree of territorial functioning among neighbourhoodyouth. Coleman (1985, p. 27) suggests, ‘graffiti can be interpreted asa way of making one’s mark in an anonymous environment thatoffers little in the way of legitimate opportunities to do so’. Whileher comment addressed the anonymity of high density towers, it isalso applicable to low density, sprawling fringe suburbs, with fewplaces or activities of interest to teenagers (Kunstler, 1996).

Other limitations relate to the study setting. Streets werepredominantly composed of new detached houses with relativelyuniformdesigns and standards ofmaintenance. Thus, data tended tobe skewed, and this necessitated using different methods todichotomise the predictor variables. Furthermore, with little timefor the social dynamic to evolve, incivilities were minimal. Thiscorrespondswithother research innewdevelopments,which foundlimited variability in objectively recorded house and yard mainte-nance (Brown, Brown, & Perkins, 2004c). Disorder tends to cluster inareas with non-residential land-uses (e.g., shops and schools)(Perkins et al., 1992), and the study developments lacked many ofthese destinations because of the lag between house constructionand the provision of facilities. While the exclusively suburbansetting limits the generalisability of the findings, the results remainof interest as they specifically reflect the suburban landscape andexperience. Nonetheless, future studies might test whether theassociations observed in this study are replicated in developmentswith greater variability in age, housing style and tenure.

Furthermore, the associationbetween themaintenancevariablesand incivilitiesmaybeconfoundedby the territorial behaviourof thehouseholders. Residents who demonstrate a sense of propriety overtheir home environment may display this same behaviour over thestreet environment, thus affecting both predictor and outcomevariables. For instance, territorial residents may remove brokenglass from public space, and report graffiti and vandalism to theirlocal council to facilitate its swift removal or repair. Although it is notpossible to identify whether territorial residents are responsible forfewer incivilities, or the local municipal services are simplyproviding a high level of service to theneighbourhood (Caughyet al.,2001). The results presented highlight a cross-sectional associationbetween the built environment and objective ratings of disorder,and therefore cannot elucidate the causal pathway. Indeed, anotherstudy limitation was the exclusive focus on physical environmentalcharacteristics, rather than combining objective datawith residents’perceptions or behaviours. Given the findings, the inclusion ofresidents’ behaviour with street-level audit data could provea fruitful line of enquiry for future research.

Finally, some items had moderate reliability, and this wouldreduce the observed magnitude of the true associations. Moderatereliabilitymayhavebeendue to the subjective natureofmanyhouseand street attributes.While the auditor’smanual attempted tomakeoperational definitions as clear as possible e an element of subjec-tivity persists. In addition, some items were highly changeable (e.g.,unkempt lawns, alcohol containers) and consequently their inter-rater reliability would have benefited if minimal time passedbetween the two audits, or preferably, if they were completedsimultaneously (Zenk et al., 2007). The NEST inter-rater audits wereconducted up to four weeks apart and this may have affectedagreement. Furthermore, inter-rater reliability could be enhancedthrough more rigorous auditor training and follow-up in the field.

5. Conclusion

This study explored the premise that house design and upkeepcould help discourage the incidence of incivilities in the publicrealm. The findings illustrate that suburban house attributes which

create opportunities for natural surveillance or reflect residents’guardianship were associated with less physical disorder in theimmediate street. In particular, the results suggest that the cumu-lative effect of several key CPTED attributes had greater potential toaffect the condition of the street environment than any singlehouse attribute. These findings underscore a connection betweenprivate and public space, where residential environments withopportunities for natural surveillance and well-defined territorymay provide visual cues as to what behaviour is appropriate ortolerated in the adjacent space. However, the relationship betweenthe built environment and incivilities is certainly more complex,and the pathway may be mediated by the social environment.Nonetheless, given the associations presented here, naturalsurveillance and territorial themes are worthy of further explora-tion as correlates of perceived safety and neighbourhood walking.

Acknowledgements

This research was funded by an Australian Research CouncilGrant (#LP0455453). The first author was initially supported by anAustralian Research Council Postgraduate Award (Industry), withthe Department for Planning and Infrastructure as industry partner,and is currently supported by an NHMRC Capacity Building Grant(#458668). The second author is supported by a NHMRC ResearchFellowship (#503712). There are no potential conflicts of interestknown to any of the contributing authors.

References

Armstrong, K., & Steinhardt, D. (2006). Understanding street racing and hoonculture: an exploratory investigation of perceptions and experience. Journal ofthe Australasian College of Road Safety, 17, 38.

Austin, D. M., Furr, L. A., & Spine, M. (2002). The effects of neighborhood conditionson perceptions of safety. Journal of Criminal Justice, 30, 417e427.

Austin, D. M., & Sanders, C. (2007). Graffiti and perceptions of safety. Journal ofCriminal Justice and Popular Culture, 14, 293e311.

Ball, K., Bauman, A., Leslie, E., & Owen, N. (2001). Perceived environmentalaesthetics and convenience and company are associated with walking forexercise among Australian adults. Preventive Medicine, 33, 434e440.

Barnes McGuire, J. (1997). The reliability and validity of a questionnaire describingneighborhood characteristics relevant to families and young children living inurban areas. Journal of Community Psychology, 25, 551e566.

Barton, H., Grant, M., & Guise, R. (2003). Shaping neighbourhoods: A guide for health,sustainability and vitality. London: Spon Press.

Boarnet, M. G., Day, K., Alfonzo, M., Forsyth, A., & Oakes, M. (2006). The Irvine-Minnesota inventory to measure built environments: reliability tests. AmericanJournal of Preventive Medicine, 30, 153.

Booth, A. (1981). The built environment as a crime deterrent. Criminology, 18,557e570.

Borst, H. C., Miedema, H. M. E., de Vries, S. I., Graham, J. M. A., & van Dongen, J. E. F.(2008). Relationships between street characteristics and perceived attractive-ness for walking reported by elderly people. Journal of EnvironmentalPsychology, 28, 353e361.

Brown, B., & Altman, I. (1983). Territoriality, defensible space and residentialburglary: an environmental analysis. Journal of Environmental Psychology, 3,203e220.

Brown, G., Brown, B., & Perkins, D. (2004c). New housing as neighbourhood revi-talization: place attachment and confidence among residents. Environment andBehavior, 36, 749e775.

Brown, B., Burton, J., & Sweaney, A. (1998). Neighbors, households, and frontporches: new urbanist community tool or mere nostalgia? Environment andBehavior, 30, 579e600.

Brown, S. C., Mason, C. A., Lombard, J. L., Martinez, F., Plater-Zyberk, E.,Spokane, A. R., et al. (2009). The relationship of built environment to perceivedsocial support and psychological distress in Hispanic elders: the role of “Eyes onthe street”. Journals of Gerontology Series B Psychological Sciences and SocialSciences, 64b, 234e246.

Brown, B., Perkins, D., & Brown, G. (2004a). Crime, new housing, and housingincivilities in a first-ring suburb: multilevel relationships across time. HousingPolicy Debate, 15, 301e345.

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

Brownson, R. C., Hoehner, C. M., Brennan, L. K., Cook, R. A., Elliott, M. B., &McMullen, K. M. (2004). Reliability of 2 instruments for auditing the environ-ment for physical activity. Journal of Physical Activity and Health, 1, 191e208.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e88 87

Brunson, L., Kuo, F. E., & Sullivan, W. C. (2001). Resident appropriation of defensiblespace in public housing: implications for safety and community. Environmentand Behavior, 33, 626e652.

Burton, E. (2005). Measuring physical characteristics of housing: the built envi-ronment site survey checklist (BESSC). Environment and Planning B: Planningand Design, 32, 265e280.

Campbell, K. (1985). Utopia found wanting. Architects Journal, 182, 57.Caughy, M. O., O’Campo, P., & Patterson, J. (2001). A brief observational measure for

urban neighborhoods. Health & Place, 7, 225e236.Cerin, E., Leslie, E., du Toit, L., Owen, N., & Frank, L. D. (2007). Destinations that

matter: associations with walking for transport. Health & Place, 13, 713e724.Coleman, A. (1985). Utopia on trial. London: Hilary Shipman Ltd.Cozens, P. M., Saville, G., & Hillier, D. (2005). Crime prevention through environ-

mental design (CPTED): a review and modern bibliography. Journal of PropertyManagement, 23, 328e356.

Craig, C. L., Brownson, R. C., Cragg, S. E., & Dunn, A. L. (2002). Exploring the effect ofthe environment on physical activity: a study examining walking to work.American Journal of Preventive Medicine, 23, 36e43.

Crowe, T. (1991). Crime prevention through environmental design. Newton, MA:Butterworth-Heinemann.

Doeksen, H. (1997). Reducing crime and the fear of crime by reclaiming NewZealand’s suburban street. Landscape and Urban Planning, 39, 243e252.

Dunstan, F., Weaver, N., Araya, R., Bell, T., Lannon, S., Lewis, G., et al. (2005). Anobservation tool to assist with the assessment of urban residential environ-ments. Journal of Environmental Psychology, 25, 293e305.

Ellaway, A., Macintyre, S., & Bonnefoy, X. (2005). Graffiti, greenery and obesity inadults: secondary analysis on European cross sectional survey. British MedicalJournal, 331(7517), 611e612.

Evans, G. W. (2003). A multimethodological analysis of cumulative risk and allo-static load among rural children. Developmental Psychology, 39, 924e933.

Evans, G. W., Kim, P., Ting, A. H., Tesher, H. B., & Shannis, D. (2007). Cumulative risk,maternal responsiveness, and allostatic load among young adolescents. Devel-opmental Psychology, 43, 341e351.

Evans, G. W., & Marcynyszyn, L. A. (2004). Environmental justice, cumulativeenvironmental risk, and health among low- and middle-income children inupstate New York. American Journal of Public Health, 94, 1942e1944.

Falconer, R., & Kingham, S. (2007). ‘Driving people crazy’: a geography of boy racersin Christchurch, New Zealand. New Zealand Geographer, 63, 181e191.

Foster, S., & Giles-Corti, B. (2008). The built environment, neighborhood crime andconstrained physical activity: an exploration of inconsistent findings. PreventiveMedicine, 47, 241e251.

Frank, L. D., Schmid, T. L., Sallis, J. F., Chapman, J., & Saelens, B. E. (2005). Linkingobjectively measured physical activity with objectively measured urban form:findings from smartraq. American Journal of Preventive Medicine, 28, 117e125.

Franzini, L., Caughy, M. O. B., Nettles, S. M., & O’Campo, P. (2008). Perceptions ofdisorder: contributions of neighborhood characteristics to subjective percep-tions of disorder. Journal of Environmental Psychology, 28, 83e93.

Giles-Corti, B., & Donovan, R. J. (2002a). The relative influence of individual, socialand physical environment determinants of physical activity. Social Science &Medicine, 54, 1793e1812.

Giles-Corti, B., & Donovan, R. J. (2002b). Socioeconomic status differences inrecreational physical activity levels and real and perceived access toa supportive physical environment. Preventive Medicine, 35, 601e611.

Giles-Corti, B., & Donovan, R. (2003). Relative influences of individual, social envi-ronmental and physical environmental correlates of walking. American Journalof Public Health, 93, 1583e1589.

Giles-Corti, B., Knuiman, M., Timperio, A., Van Niel, K., Pikora, T. J., Bull, F. C. L.,et al. (2008). Evaluation of the implementation of a state governmentcommunity design policy aimed at increasing local walking: design issues andbaseline results from RESIDE, Perth Western Australia. Preventive Medicine, 46,46e54.

Greenberg, S. W., Rohe, W. M., & Williams, J. R. (1982). Safety in urban neighbor-hoods: a comparison of physical characteristics and informal territorial controlin high and low crime neighborhoods. Population and Environment, 5, 141e165.

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

Hillier, B. (1986). City of Alice’s dreams. Architects Journal. 9 July.Jackson, J. B. (1978). Front yards. In S. Kaplan, & R. Kaplan (Eds.), Humanscape:

Environments for people. North Scituate: Duxbury Press.King, D. (2008). Neighborhood and individual factors in activity in older adults:

results from the neighborhood and senior health study. Journal of Aging andPhysical Activity, 16, 144e170.

Kunstler, J. H. (1996). Home from nowhere. New York: Simon and Schuster.Kuo, F. E., & Sullivan, W. C. (2001). Environment and crime in the inner city. Envi-

ronment and Behavior, 33, 343e367.Lee, R. E., Booth, K. M., Reese-Smith, J. Y., Regan, G., & Howard, H. H. (2005). The

physical activity resource assessment (PARA) instrument: evaluating features,amenities and incivilities of physical activity resources in urban neighborhoods.International Journal of Behavioral Nutrition and Physical Activity, 2.

Loukaitou-Sideris, A., & Eck, J. E. (2007). Crime prevention and active living.American Journal of Health Promotion, 21, 380e389.

McCormack, G. R., Giles-Corti, B., & Bulsara, M. (2008). The relationship betweendestination proximity, destination mix and physical activity behaviors.Preventive Medicine, 46, 33e40.

Mendes de Leon, C. F., Cagney, K. A., Bienias, J. L., Barnes, L. L., Skarupski, K. A.,Scherr, P. A., et al. (2009). Neighborhood social cohesion and disorder in relationto walking in community-dwelling older adults: a multilevel analysis. Journal ofAging and Health, 21, 155e171.

Merry, S. E. (1981). Defensible space undefended: social factors in crime controlthrough environmental design. Urban Affairs Quarterly, 16, 397e422.

Michael, Y. L., Green, M. K., & Farquhar, S. A. (2006). Neighborhood design and activeaging. Health & Place, 12, 734e740.

Miles, R. (2008). Neighborhood disorder, perceived safety, and readiness toencourage use of local playgrounds. American Journal of Preventive Medicine, 34,275e281.

Minnery, J. R. (2005). Measuring crime prevention through environmental design.Journal of Architecture and Planning Research, 22, 330e341.

Nagel, C. L., Carlson, N. E., Bosworth, M., & Michael, Y. L. (2008). The relationbetween neighborhood built environment and walking activity among olderadults. American Journal of Epidemiology, 168, 461.

Nassauer, J. I., Wang, Z., & Dayrell, E. (2009). What will the neighbours think?Cultural norms and ecological design. Landscape and Urban Planning, 92,282e292.

Owen, N., Cerin, E., Leslie, E., duToit, L., Coffee, N., Frank, L. D., et al. (2007).Neighborhood walkability and the walking behavior of Australian adults.American Journal of Preventive Medicine, 33, 387e395.

Owen, N., Humpel, N., Leslie, E., Bauman, A., & Sallis, J. F. (2004). Understandingenvironmental influences on walking: review and research agenda. AmericanJournal of Preventive Medicine, 27, 67e76.

Perkins, D. D., Florin, P., Rich, R. C., Wandersman, A., & Chavis, D. M. (1990).Participation and the social and physical environment of residential blocks:crime and community context. American Journal of Community Psychology, 18,83e115.

Perkins, D. D., Meeks, J. W., & Taylor, R. B. (1992). The physical environment of streetblocks and resident perceptions of crime and disorder: implications for theoryand measurement. Journal of Environmental Psychology, 12, 21e34.

Phillips, T., & Smith, P. (2006). Rethinking urban incivility research: strangers,bodies and circulations. Urban Studies, 43, 879e901.

Pikora, T. J., Bull, F. C. L., Jamrozik, K., Knuiman, M., Giles-Corti, B., & Donovan, R. J.(2002). Developing a reliable audit instrument to measure the physical envi-ronment for physical activity. American Journal of Preventive Medicine, 23,187e194.

Pikora, T. J., Giles-Corti, B., Knuiman, M. W., Bull, F., Jamrozik, K., & Donovan, R. J.(2006). Neighborhood environmental factors correlated with walking nearhome: using spaces. Medicine and Science in Sports and Exercise, 38,708e714.

Raco, M. (2007). Securing sustainable communities: citizenship, safety andsustainability in the new urban planning. European Urban and Regional Studies,14, 305e320.

Raudenbush, S. W., & Sampson, R. J. (1999). Ecometrics: towards a science ofassessing ecological settings, with application to the systematic social obser-vation of neighborhoods. Sociological Methodology, 29, 1e41.

Reisig, M. D., & Cancino, J. M. (2004). Incivilities in non metropolitan communities:the effects of structural constraints, social conditions, and crime. Journal ofCriminal Justice, 32, 15e29.

Robinson, J., Lawton, B., Taylor, R. B., & Perkins, D. D. (2003). Multilevel longitudinalimpacts of incivilities: fear of crime, expected safety and block satisfaction.Journal of Quantitative Criminology, 19, 237e274.

Ross, C. (1993). Fear of victimization and health. Journal of Quantitative Criminology,9, 159e175.

Ross, C., Mirowsky, J., & Pribesh, S. (2002). Disadvantage, disorder, and urbanmistrust. City & Community, 1, 59e82.

Schneider, R. H., & Kitchen, T. (2007). Crime prevention and the built environment.New York: Routledge.

Schweitzer, J. H., Kim, J. W., & Macklin, J. R. (1999). The impact of the built envi-ronment of crime and fear of crime in urban neighborhoods. Journal of UrbanTechnology, 6, 59e73.

Shenassa, E. D., Leibhaber, A., & Ezeamama, A. (2006). Perceived safety of area ofresidence and exercise: a pan-European study. American Journal of Epidemi-ology, 163, 1012e1017.

Skogan, W. (1990). Disorder and decline: Crime and the spiral of decay in Americanneighborhoods. New York: The Free Press.

Smardon, R. C. (1988). Perception and aesthetics of the urban environment: reviewof the role of vegetation. Landscape and Urban Planning, 15, 85e106.

Southworth, M., & Owens, P. (1993). The evolving metropolis. Journal of the Amer-ican Planning Association, 59, 271e288.

Stokols, D. (1992). Establishing and maintaining health environments. AmericanPsychologist, 47, 6e22.

Sugiyama, T., Leslie, E., Giles-Corti, B., & Owen, N. (2009). Physical activity forrecreation or exercise on neighbourhood streets: associations with perceivedenvironmental attributes. Health & Place, 2009, 1058e1063.

Sugiyama, T., & Ward Thompson, C. (2008). Associations between characteristics ofneighbourhood open space and older people’s walking. Urban Forestry & UrbanGreening, 7, 41e51.

Taylor, R. B. (1987). Toward an environmental psychology of disorder: delinquency,crime and fear of crime. In D. Stokols, & I. Altman (Eds.), Handbook of envi-ronmental psychology, Vol. 2 (pp. 951e986). New York: Wiley.

Taylor, R. B. (1988). Human territorial functioning. Cambridge: Cambridge UniversityPress.

S. Foster et al. / Journal of Environmental Psychology 31 (2011) 79e8888

Taylor, R. B., Gottfredson, S. D., & Brower, S. (1980). The defensibility of defensiblespace. In T. Hirschi, & S. D. Gottfredson (Eds.), Understanding crime: Currenttheory and research. California: Sage.

Taylor, R. B., Koons, B., Kurtz, E., Greene, J., & Perkins, D. D. (1995). Street blocks withmore non-residential land use have more physical deterioration. Urban AffairsReview, 31, 120e136.

Werner, C., Peterson-Lewis, S., & Brown, B. (1989). Inferences about homeowners’sociability: impact of Christmas decorations and other cues. Journal of Envi-ronmental Psychology, 9, 279e296.

Winchester, S., & Jackson, H. (1982). Residential burglary: The limits of prevention (No.Home Office Research Study No. 74). London: Her Majesty’s Stationery Office.

Wood, L., & Giles-Corti, B. (2008). Is there a place for social capital in the psychologyof health and place? Journal of Environmental Psychology, 28, 154e163.

Wood, L., Shannon, T., Bulsara, M., Pikora, T., McCormack, G., & Giles-Corti, B. (2008).The anatomy of the safe and social suburb: an exploratory study of the builtenvironment, social capital and residents’ perceptions of safety. Health & Place,14, 15e31.

Zelinka, A., & Brennan, D. (2001). Safescape. Chicago: American PlanningAssociation.

Zenk, S. N., Schulz, A. J., Mentz, G., House, J. S., Gravlee, C. C., Miranda, P. Y., et al.(2007). Inter-rater and test-retest reliability: methods and results for theneighborhood observational checklist. Health & Place, 13, 452e465.