Accepted Manuscript Are Liveable Neighbourhoods Safer Neighbourhoods? Testing the Rhetoric on New Urbanism and Safety from Crime in Perth, Western Australia Sarah Foster, Paula Hooper, Matthew Knuiman, Fiona Bull, Billie Giles-Corti PII:

S0277-9536(15)00242-7

DOI:

10.1016/j.socscimed.2015.04.013

Reference:

SSM 10043

To appear in:

Social Science & Medicine

Please cite this article as: Foster, S., Hooper, P., Knuiman, M., Bull, F., Giles-Corti, B., Are Liveable Neighbourhoods Safer Neighbourhoods? Testing the Rhetoric on New Urbanism and Safety from Crime in Perth, Western Australia, Social Science & Medicine (2015), doi: 10.1016/j.socscimed.2015.04.013. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Are Liveable Neighbourhoods Safer Neighbourhoods? Testing the Rhetoric on New Urbanism and Safety from Crime in Perth, Western Australia

RI PT

Sarah Fostera*, Paula Hoopera, Matthew Knuimanb, Fiona Bulla and Billie Giles-Cortic

a

Centre for the Built Environment and Health, School of Sport Science, Exercise & Health and

School of Earth & Environment, The University of Western Australia (M707), 35 Stirling Highway,

SC

Crawley WA 6009, Australia Email: [email protected]

M AN U

Telephone: +61 8 6488 8730

b

School of Population Health, The University of Western Australia, 35 Stirling Highway, Crawley

TE D

WA 6009, Australia

c

McCaughey VicHealth Centre for Community Wellbeing, Melbourne School of Population Health,

EP

University of Melbourne Australia

AC C

*Corresponding Author

Acknowledgements

RESIDE was funded by grants from the Western Australian Health Promotion Foundation (Healthway) (#11828), the Australian Research Council (ARC) (#LP0455453) and supported by an Australian National Health & Medical Research Council (NHMRC) Capacity Building Grant (#458688). The first author is supported by a Healthway Health Promotion Research Fellowship (#21363); the second author by a NHRMC CRE in Healthy Liveable Communities postdoctoral

ACCEPTED MANUSCRIPT

fellowship (#1061404); and the last author by a NHMRC Principal Research Fellowship (#1004900). Nick Middleton, Sharyn Hickey, Bridget Beasley and Dr Bryan Boruff are gratefully acknowledged for their assistance and advice in the development of the GIS measures in this study, and The Western Australian Land Information Authority (©2003), Western Australian Department of

RI PT

Planning for provided the core spatial data. Crime locations were supplied courtesy of the

AC C

EP

TE D

M AN U

SC

Western Australia Police.

ACCEPTED MANUSCRIPT

1

Are Liveable Neighbourhoods Safer Neighbourhoods? Testing the Rhetoric on New Urbanism

2

and Safety from Crime in Perth, Western Australia

3

Abstract

5

New urbanism advocates for the design of the compact, pedestrian-friendly, mixed-use

6

developments thought to promote walking. New urbanist proponents also claim their

7

developments incur other social and wellbeing benefits, including enhanced safety from crime;

8

however there is limited empirical evidence supporting this. We tested the premise that new

9

urbanism inhibits crime by examining the relationship between compliance with a planning policy

M AN U

SC

RI PT

4

based on new urbanism and: (1) residents’ reports of victimisation; and (2) objective crime

11

measures. RESIDE Participants (n=603) who had lived in their new developments for 36 months

12

completed a questionnaire that included items on their experiences of victimisation. Detailed

13

measures quantifying the degree to which these developments (n=36) complied with the policy

14

requirements were generated in Geographic Information Systems. Logistic regression examined

15

the associations between policy compliance and self-report victimisation, and negative binomial

16

log-linear models examined area-level associations between compliance and objective crime. For

17

each 10% increase in overall policy compliance, the odds of being a victim reduced by 40%

18

(OR=0.60, CI=0.53-0.67, p=0.000). Findings for the individual policy ‘elements’ were consistent

19

with this: for each 10% increment in compliance with the community design, movement network,

20

lot layout and public parkland elements, the odds of victimisation reduced by approximately 6%

21

(p=0.264), 51% (p=0.001), 15% (p=0.000) and 22% (p=0.001) respectively. However, while policy

22

compliance correlated with lower odds of self-report victimisation among residents, the

23

associations between compliance and development-wide (objective) crime were positive but non-

24

significant. The results indicate that planning policies based on new urbanism may indeed deliver

AC C

EP

TE D

10

1

ACCEPTED MANUSCRIPT

25

other social and wellbeing benefits for residents, however they also hint that the design of an

26

‘objectively’ safe place may differ from the design of a ‘subjectively’ safe space.

27

Key Words: New urbanism; victimisation; crime; safety; planning policy; built environment

RI PT

28 29

Introduction

31

New urbanism advocates for the design of the compact, pedestrian-friendly, mixed-use

32

developments thought to promote walking, minimise car dependence and enhance sense of

33

community (Congress for the New Urbanism, 2001; Duany et al., 2000). To some extent these

34

claims are supported – the accumulated evidence suggests that developments designed in

35

accordance with new urbanism principles can positively impact residents walking behaviours (Dill,

36

2006; Hooper et al., 2014; Lund, 2003; Rodríguez et al., 2006), and even facilitate social contact

37

between residents (Kim & Kaplan, 2004; Leyden, 2003; Talen & Koschinsky, 2014). Proponents of

38

new urbanism also claim their developments incur other social benefits for residents, including

39

enhanced safety from crime (Duany et al., 2000) however to date; there is little empirical evidence

40

to support this assertion (Cozens, 2008; Cozens & Hillier, 2012; Schneider & Kitchen, 2007).

M AN U

TE D

EP

AC C

41

SC

30

42

There are two competing perspectives on the notion that new urbanism creates safer

43

neighbourhoods. On one hand, urban planners argue that mixed-use neighbourhoods generate

44

more pedestrian traffic, making streets safer through natural surveillance or ‘eyes on the street’

45

(Duany et al., 2000; Jacobs, 1961; Zelinka & Brennan, 2001). Jacobs (1961) envisaged that

46

pedestrians would make streets lively and interesting to watch, and in turn this would encourage

47

further surveillance from adjacent buildings. This approach is embedded within new urbanism,

48

despite Jacobs’ caution against transferring her ideas to suburban settings. For example, the 2

ACCEPTED MANUSCRIPT

49

Charter for the New Urbanism emphasises that ‘streets and squares should be safe, comfortable

50

and interesting to the pedestrian’, and that by adhering to new urbanism principles (i.e., if places

51

are ‘properly configured’) ‘they encourage walking and enable neighbours to know each other and

52

protect their communities’ (Congress for the New Urbanism, 2001).

RI PT

53

Conversely, evidence from criminology links key elements of new urbanism with increased crime

55

levels (Cozens, 2008; Schneider & Kitchen, 2007). For example, the non-residential land uses that

56

provide destinations to walk to (e.g., shopping centres, recreational facilities and transport nodes)

57

have been associated with higher levels of property crime (Beavon et al., 1994; Bowes, 2007;

58

Brantingham & Brantingham, 1993; McCord et al., 2007), and the presence of drinking venues and

59

alcohol sales linked with more violent crime (Gorman et al., 2001; Gruenewald et al., 2006; Popova

60

et al., 2009). Similarly, street connectivity is integral to new urbanism as it provides both direct

61

and varied walking routes for residents. However, better connected streets (i.e., gridded street

62

layouts) are also more easily navigated by would-be offenders, with more potential ‘escape

63

routes’ (Brantingham & Brantingham, 1993). Indeed, in the criminology literature there is general

64

consensus that higher street connectivity increases vulnerability to crime (Cozens, 2008; Cozens &

65

Hillier, 2008; Schneider & Kitchen, 2007). Still, it is also worth noting that the individual attributes

66

of a compact, walkable neighbourhood rarely exist in isolation (Sallis et al., 2012), and their

67

cumulative presence may be most pertinent to crime. For example, permeable streets may not

68

impact crime unless destinations are present that attract potential offenders (Brantingham &

69

Brantingham, 1993).

AC C

EP

TE D

M AN U

SC

54

70 71

Despite these differing perspectives on what design features constitute a safer neighbourhood,

72

few empirical studies have examined the relationship between new urbanism and crime (Knowles,

73

2003) and rarer still are studies that assess the degree of new urban policy implementation and 3

ACCEPTED MANUSCRIPT

74

residents’ experiences of crime. This study addresses this evidence gap and tests the premise that

75

new urbanism can enhance personal safety by examining the relationship between compliance

76

with a planning policy based on new urbanist principles and: (1) residents’ experiences of

77

victimisation; and (2) crimes committed within the development.

RI PT

78

Methods

80

Study context

81

In February 1998 the Western Australian State Government began trialling the ‘Liveable

82

Neighbourhoods Community Design Guidelines’ (LN). This was introduced to replace the

83

conventional design codes that had facilitated car dependence and sprawl, and stimulate the

84

development of sustainable suburban communities (Western Australian Planning Commission,

85

2000). LN is essentially a local interpretation of New Urbanism, tailored to the Western Australian

86

context and in 2001 was recognised with an annual charter award from the Congress of New

87

Urbanism (Congress for the New Urbanism, 2007).

M AN U

TE D

88

SC

79

The second edition of the guidelines published in 2000 consisted of six general design topics,

90

termed ‘elements’ (Western Australian Planning Commission, 2000). Four of these elements

91

(community design; movement networks; lot layout; public parkland) aimed to provide an

92

alternative approach to the design of (suburban) neighbourhoods by creating more compact,

93

pedestrian-friendly neighbourhoods, with good links to public transport services. The key

94

intended outcomes of the LN policy were to reduce car dependence and encourage more active

95

forms of transport; however the policy also aimed to enhance personal safety, primarily through

96

increased surveillance and activity (Western Australian Planning Commission, 2000).

AC C

EP

89

97 98

Measuring implementation of policy requirements 4

ACCEPTED MANUSCRIPT

A process evaluation was undertaken to objectively measure the on-ground implementation of 43

100

unique, quantifiable requirements from the LN policy across its four elements (community design;

101

movement network; lot layout; public parkland using Geographic Information Systems (GIS).

102

These detailed measures were created for 36 new housing developments being built in Perth,

103

Western Australia – 19 of which purported to be developed according to LN and 17

104

‘conventionally designed’ developments that matched these LN developments on size and location

105

(i.e., distance from the ocean) (Hooper et al., 2014). The timing of the LN evaluation was chosen

106

to coincide with the third time-point (i.e., 36 months) of the RESIDential Environments (RESIDE)

107

Project, a longitudinal natural experiment evaluating the impact of LN. Full details of the process

108

evaluation methods and the development of measures in GIS for each of these policy

109

requirements are reported elsewhere (Hooper et al., 2014). The four elements are briefly

110

summarised below.

M AN U

SC

RI PT

99

111

Community Design: sets out objectives for designing ‘complete integrated communities’, rather

113

than the segregated residential (or dormitory) developments typical of conventionally planned

114

suburbs (Western Australian Planning Commission, 2000). A key community design principle

115

relates to the configuration of the neighbourhood and town centres, with an emphasis on creating

116

more traditional main-street mixed-use centres where pedestrian-scaled, street-fronting retail

117

layouts predominate. LN aims to create hubs of destinations with sufficient diversity to be useful

118

walkable nodes and which act as community focal points that attract people for a variety of

119

activities (Western Australian Planning Commission, 2000).

AC C

EP

TE D

112

120 121

Movement Network: LN advocates for a highly interconnected street system with good internal

122

and external access aimed at reducing local travel distances and optimising walkable access to

123

centres, schools, public transport and other destinations. The policy specifies standards for block 5

ACCEPTED MANUSCRIPT

sizes that create a more connected street network, based on the premise that smaller (or shorter)

125

blocks create a denser network of streets, and reduce walking distances and increase route

126

choices between locations (Dill, 2003; Song, 2005). Further, while the policy recognises cul-de-sac

127

are a popular suburban street pattern, and as such does not prohibit their use, it applies standards

128

to help limit their use and ensure they do not impede the overall connectivity of the movement

129

network. These standards include specifications relating to cul-de-sac length, the provision of

130

linking routes at the terminating point of the cul-de-sac, the number of residential lots that should

131

be positioned on cul-de-sac and the total proportion of residential dwellings within the

132

development that should be situated on these ‘dead-end streets’.

SC

RI PT

124

M AN U

133

Lot Layout: This element places an emphasis on creating greater residential densities and the

135

provision of a mixture of lot sizes distributed throughout the developments to facilitate housing

136

variety, choice and affordability, and to cater for increasingly diverse household types. Standards

137

for the provision of smaller lots and locating of lots for mixing of compatible uses near centres and

138

public transport stops are stipulated to achieve sufficient densities to support these businesses

139

and services.

EP

140

TE D

134

Public Parkland: LN requires a minimum contribution of 10% of the gross subdivisible land area in

142

new developments be provided as public parkland and identifies three different park types based

143

on size and catchment areas to provide for a range of uses and activities: local parks,

144

neighbourhood parks and district parks. The policy seeks to provide a range of parkland types

145

which are safe and conveniently located for the majority of residents. Under LN schools are also

146

encouraged to share their ovals or playing fields as community facilities for out of school hour’s

147

use. Unlike current conventional development practices, LN requires that virtually all public

148

parkland be overlooked by development, rather than being backed onto by development.

AC C

141

6

ACCEPTED MANUSCRIPT

149

Quantifying policy compliance

151

Policy compliance was defined as the degree to which the construction of the developments

152

adhered with, or met, the standards outlined by LN. A simple scoring system was developed to

153

quantify the extent to which the 43 measureable requirements had been implemented as

154

intended by the LN (Hooper et al., 2014). The level of compliance for each element and overall

155

was calculated as the percentage of the maximum policy implementation score

156

attainable/intended.

SC

RI PT

150

157

Participant data

159

The sample comprised a subset of participants from the larger RESIDE Project. Briefly, all people

160

building homes in one of 73 new developments were invited to participate (response rate 33.4%).

161

Once recruited, participants completed a questionnaire before they moved into their new home,

162

and on three subsequent occasions after they relocated (at 12, 36 and 84 months). Full details are

163

described elsewhere (Giles-Corti et al., 2008). RESIDE was approved by The University of Western

164

Australia’s Human Research Ethics Committee.

TE D

EP

165

M AN U

158

While 1812 participants completed the initial baseline survey, 1220 participants (67%) remained in

167

the study at 36 months. This cross-sectional study focused on the subset of these RESIDE

168

participants (n=603) who completed the 36 month survey and continued to live in one of the 36

169

developments that were assessed for compliance with the LN policy. Of the 36 developments

170

included in this study, approximately 90% were located in new Greenfield developments on the

171

urban fringe. These developments were characterised by expanses of single family detached

172

houses with good access to public open space and walking infrastructure (Hooper et al., 2014) but

173

relatively poor access to shops and services (Christian et al., 2013; Foster et al., 2010). This

AC C

166

7

ACCEPTED MANUSCRIPT

pattern is typical of new suburban developments in Perth (Hooper et al., 2014). Australian Bureau

175

of Statistics (ABS) rankings also indicated the study developments were located in areas with

176

relatively low levels of socio-economic disadvantage. Furthermore, compared to the wider Perth

177

metropolitan area population, our study sample was slightly older, more likely to be married and

178

more affluent (Australian Bureau of Statistics, 2007), reflecting a population group able to access

179

finance and purchase a new home.

RI PT

174

180

The RESIDE questionnaire collected information on a range of demographic variables, including

182

participant age, gender, education, marital status and number of children. Self-reported

183

victimisation was measured by asking participants whether they, or anyone they personally knew,

184

had been the victim the following crimes in their neighbourhood in the last two years: (1)

185

household burglary; (2) harassed or threatened while in public; and (3) physically attacked or

186

mugged while in public. ‘Objective’ crime location data for the year matching survey completion

187

were supplied by the Western Australian Police. Crime measures (counts) were created for

188

offences committed within the extents of each of the study development boundaries, focusing on:

189

(1) crimes committed against the person in public space (e.g., threats, disorderly behaviour,

190

assault; robbery); and (2) actual and attempted burglaries.

EP

TE D

M AN U

SC

181

AC C

191 192

Statistical analysis

193

All analyses were conducted in SPSS version 22. First, separate logistic regression models were

194

run with generalised estimating equations (GEE) to account for clustering within residential

195

development to examine associations between compliance with each of the LN elements, and

196

overall policy compliance, and self-report victimisation (binary dependent variable: yes/no).

197

Estimated odds ratios from these models represent the increase in odds of self-reported

198

victimisation for each 10% increase in the level of compliance with the LN policy (Table 2). All 8

ACCEPTED MANUSCRIPT

199

models adjusted for age, gender, education, marital status, number of children living at home,

200

area socio-economic status (IRSD), stage of construction and size of development.

201

Next, a series of backwards stepwise elimination models were run separately for the LN

203

requirements within each of the four elements to identify the specific requirements from each

204

element that were most strongly associated with the victimisation outcome. At each stage, the LN

205

requirement with the highest p value (and >0.05) was removed and the model re-fitted. This

206

continued until all remaining LN requirements had p values ≤0.05 which constituted the (final)

207

multivariate model (Table 3). Again, all models controlled for age, gender, education, marital

208

status, number of children living at home, area socio-economic status (IRSD), stage of construction

209

and size of development.

M AN U

SC

RI PT

202

210

Finally, negative binomial log-linear models examined the area-level associations between LN

212

compliance (by element and overall) and number of crimes reported to police occurring within the

213

extents of the development (Table 4). These models adjusted for area socio-economic status

214

(IRSD), stage of construction and size of development. The log of the number of dwellings in the

215

development was included as an offset term so that the estimated effects of compliance scores

216

were on the crime rate per dwelling.

217

AC C

EP

TE D

211

218

Results

219

Approximately 29% of participants reported some form of victimisation in the past two years

220

within their local neighbourhood (Table 1). Table 2 presents the (adjusted) odds ratios for the

221

associations between LN policy compliance and victimisation. For each 10% increase in the total

222

level of compliance, the odds of being a victim reduced by 40% (OR=0.60, CI=0.53-0.67, p=0.000).

223

Findings for the individual LN elements were consistent with this overall finding. For each 10% 9

ACCEPTED MANUSCRIPT

224

increment in compliance with the community design, movement network, lot layout and public

225

parkland elements, the odds of victimisation reduced by approximately 6% (p=0.264), 51%

226

(p=0.001), 15% (p=0.000) and 22% (p=0.001) respectively.

227

Table 3 reports the specific requirements (of the 43 that were measured) within each LN element

229

that were most strongly associated with self-reported victimisation are shown in Table 3. While

230

compliance with the community design element (outlined above) was not significantly associated

231

with victimisation, the configuration of the neighbourhood centre appears important. Compared

232

to participants with no community centre, those with a ‘big box’ centre configuration were more

233

likely to report victimisation (OR=1.42, p=0.031), whereas there was no association between

234

having a main street configured centre and victimisation (p=0.676).

M AN U

SC

RI PT

228

235

Two movement network requirements were associated with reduced odds of victimisation. A

237

higher sidewalk to road ratio (where higher values indicate more roads with an adjacent sidewalk)

238

and higher tree density along footpaths were associated with reduced odds of victimisation

239

(p=0.017 and p=0.010 respectively). The effect size for tree density along footpaths was

240

considerable (OR=0.367, CI=0.17-0.78); however this requirement encapsulated the specificity of

241

presence of footpaths and trees (with these trees positioned along the footpath).

EP

AC C

242

TE D

236

243

Several lot layout requirements were associated with victimisation. While the overall measure of

244

the net residential dwelling density in the development was associated with increased

245

victimisation (OR=1.24, p=0.000), other lot layout requirements suggest that the mix of housing

246

available may be protective against victimisation. Notably, areas with more diverse housing

247

indicated by the percentage of residential lots less than 350m2 (i.e., relatively small, cottage lots)

248

and the number of different lot sizes present were both associated with reduced odds of 10

ACCEPTED MANUSCRIPT

249

victimisation (p=0.031 and p=0.018 respectively). Conversely larger mean lot sizes (a measure

250

reflecting a lack of housing diversity) were associated with increased odds of victimisation

251

(p=0.000).

252

The public parkland element requirements stipulating proximate access to different park types

254

were consistently associated with reduced odds of victimisation. A higher percentage of houses

255

with a park within 400m was associated with reduced odds of victimisation (p=0.000). This

256

pattern was consistent for both small (0.3 ≤ 0.5 ha) and large neighbourhood parks (1.5 ≤ 2.5 ha).

SC

RI PT

253

257

Finally, we tested the development-level associations between LN policy compliance and objective

259

measures of crime within the development (Table 4). There were no significant associations

260

between LN compliance (overall, or by element) and actual or attempted burglary, or crimes

261

committed against the person in public space.

262

TE D

M AN U

258

Discussion

264

New urbanism is frequently espoused as generating multiple community benefits, including the

265

promotion of walking, public transport use and sense of community, and even enhanced

266

community safety (Duany et al., 2000). However, the purported social benefits of new urbanism

267

are rarely substantiated in evidence. This study provides empirical evidence that new urbanist

268

design has the potential to discourage neighbourhood crime. We tested the impact of compliance

269

with a new urbanism inspired planning policy on residents’ experiences of victimisation, and found

270

that self-report victimisation reduced with increasing overall policy compliance. This pattern was

271

consistent for the four policy ‘elements’ that comprised the overall LN compliance score (albeit

272

non-significant for the ‘community design’ element). Furthermore, within each LN ‘element’,

273

several individual requirements were significantly associated with victimisation.

AC C

EP

263

11

ACCEPTED MANUSCRIPT

274

The LN ‘community design’ element emphasises the creation of more traditional main-street

276

mixed-use centres where pedestrian-scaled, street-fronting retail layouts predominate (Western

277

Australian Planning Commission, 2000). Notably, the ‘community design’ requirements relating to

278

the town centre configuration were associated with victimisation, with participants in

279

developments served by a ‘big box’ centre more likely to report being a victim. In this context, ‘big

280

box’ neighbourhood centres were characterised by clusters of shops and services that

281

accommodated the day-to-day needs of local residents, typically surrounded by a generous car

282

park provision. These areas can be hostile or intimidating landscapes for pedestrians and cyclists

283

as there are often insufficient footpaths, marked crossing areas and traffic controls. Pedestrians

284

are often forced to navigate through vast expanses of parking where cars have ‘right of way’

285

(Falconer, 2008). By contrast, the main street configured centres position parking at the rear of

286

buildings, preserving the connection between the building and street with increased potential for

287

natural surveillance. Jacobs (1961) theorised that business proprietors would act as ‘sidewalk

288

guardians’ – monitoring interactions and intervening at signs of trouble (Jacobs, 1961). Further,

289

pedestrian oriented main streets are thought to encourage social interactions between local

290

residents, and have been linked to greater sense of community (Pendola & Gen, 2008) and social

291

capital (Wood et al., 2010) - constructs which correlate with enhanced feelings of safety (Foster et

292

al., 2010; Wood et al., 2008). For example, Wood et al. (2010) observed a positive association

293

between commercial floor area ratio (i.e., a measure of walkable site design where higher values

294

indicate less surface area for parking, with shops and services positioned closer to the sidewalk)

295

and sense of community (Wood et al., 2010).

AC C

EP

TE D

M AN U

SC

RI PT

275

296 297

Two individual ‘movement network’ requirements were significantly (negatively) associated with

298

victimisation: the proportion of streets with adjacent sidewalks and tree density along footpaths. 12

ACCEPTED MANUSCRIPT

Both requirements contribute to a superior environment for pedestrians, and have been found to

300

be associated with higher levels of walking among residents (Hooper et al., 2014), potentially

301

increasing natural surveillance. Our finding relating to the presence of street trees and

302

victimisation warrants further discussion. To date, there is mixed evidence on the relationship

303

between vegetation and crime. Vegetation can conceal perpetrators as they select a target,

304

commit an offence and flee the scene (Nasar et al., 1993), which is thought to promote fear by

305

limiting visibility in the immediate vicinity (Nasar & Jones, 1997). However, in residential settings,

306

vegetation has been associated with less fear of crime (Nasar, 1982), a greater sense of safety

307

among residents (Kuo et al., 1998; Maas et al., 2009) and lower reported crime (Kuo & Sullivan,

308

2001b). Indeed, Donovan and Prestemon (2012) suggest the type and location of vegetation may

309

account for conflicting findings (Donovan & Prestemon, 2012). They found that smaller trees on

310

residential lots (i.e., that impede visibility) were associated with increased crime, whereas street

311

trees (or larger private trees) were associated with lower crime rates. Further, while street trees

312

contribute to a more pleasant environment (Cervero & Kockelman, 1997) they also act to slow

313

traffic speeds, increasing the safety of the street for pedestrians and cyclists (Burden, 2006; City of

314

Melbourne, 2011; Western Australian Planning Commission, 2009) and affording more effective

315

natural surveillance from vehicles.

SC

M AN U

TE D

EP

AC C

316

RI PT

299

317

This connection between vegetation and victimisation was also confirmed by the ‘public parkland’

318

element results. LN aims to ensure that the design of development surrounding parks provides

319

high levels of visual supervision by residents to enhance personal and property security,

320

deterrence of crime and vandalism, and promotion of safety for park users (Western Australian

321

Planning Commission, 2000). Three specific requirements were negatively associated with

322

victimisation, and all related to the proportion of dwellings within 400m of a park. While there

323

were some significant results based on park size (i.e., local or large neighbourhood park), the 13

ACCEPTED MANUSCRIPT

effect size was greatest for the proportion of dwellings within 400m of ‘any park’. This is

325

consistent with a recent evidence review that concluded: ‘the greener the residential setting, the

326

safer it is perceived’ (Maruthaveeran & van den Bosch, 2014 p.13). Indeed, our study emphasised

327

the relevance of both park access and street trees to residents’ reports of victimisation. Several

328

potential pathways could explain this connection between green vegetation/space, and

329

victimisation: (1) an aesthetically pleasing public realm might draw residents into the streets and

330

parks, enhancing natural surveillance; (2) visually appealing streetscapes might attract further

331

surveillance from surrounding houses, regardless of whether residents are ‘people watching’; (3)

332

green vegetation can create a calming restorative environment that helps alleviate feelings of

333

anger, frustration and aggression (Kuo & Sullivan, 2001a); and (4) attractive public spaces provide

334

settings where people can gather, interact and develop social ties (Bedimo-Rung et al., 2005;

335

Chiesura, 2004; Kuo et al., 1998; Wolch et al., 2014), potentially enhancing safety. Indeed, a

336

comparison of new urbanist and traditional developments attributed the greater sense of

337

community in new urbanist communities to more natural features and shared spaces (Kim &

338

Kaplan, 2004).

SC

M AN U

TE D

EP

339

RI PT

324

The final LN element focuses on ‘lot layout’ and includes requirements relating to lot size, housing

341

diversity and residential density. While development-wide residential density was positively

342

associated with being a victim, the other lot layout findings suggested that, independent of overall

343

density, the design and form of density could mitigate reports of victimisation. For instance,

344

requirements relating to the provision of smaller lots (i.e., less than 350m2) and a greater diversity

345

of lot sizes were significantly associated with reduced odds of victimisation. Further, as mean lot

346

sizes increased within the development, so too did self-report victimisation. On balance, these

347

results indicate that, while density per se may increase the odds of victimisation, the design of

348

residential lots and range of housing stock available in the development may enhance safety. In

AC C

340

14

ACCEPTED MANUSCRIPT

this primarily residential (suburban) setting, small lots are situated closer to the street (i.e., smaller

350

setbacks), increasing the visual connection between the residence and street. Indeed, LN

351

recognises that the height, character and visual permeability of lot boundaries and fences impacts

352

the potential for surveillance from the building over the park or street. The policy also aims to

353

provide lots without street frontages being dominated by garages and parked cars (Western

354

Australian Planning Commission, 2000). This stands in contrast to the conventional style of

355

residential construction where the dwelling is typically set back from the street on a large lot,

356

served by a drive way and often behind, or obscured by, a garage or car-port.

SC

RI PT

349

357

The notion that house design can help augment safety is supported by other research in the same

359

developments which indicated that house designs that created opportunities for natural

360

surveillance (e.g., visible windows, verandahs) contributed to less graffiti and disorder in the street

361

(Foster et al., 2011). Furthermore, by providing diverse housing options, developments cater to a

362

broader population (e.g., young people, retirees, mixed incomes), where residents’ different daily

363

schedules and use of space ensure the more continuous presence of ‘eyes on the street’ or

364

guardians monitoring the neighbourhood over different time-frames (Jacobs, 1961). The charter

365

of new urbanism states that ‘within neighbourhoods a broad range of housing types and price

366

levels can bring people of diverse ages, races and incomes into daily interaction, strengthening the

367

personal and civic bonds essential to an authentic community’ (Congress for the New Urbanism,

368

2001). To some extent, our findings indicate that adherence to the urbanism planning principles

369

intended to enhance social equity (Talen, 2002) could produce co- benefits, including community

370

safety.

AC C

EP

TE D

M AN U

358

371 372

Our study examined associations between LN and both residents’ reports of victimisation and

373

development-level crime (i.e., crimes reported to police). Results indicate that the 15

ACCEPTED MANUSCRIPT

implementation of LN impacts residents’ experiences of crime, or at the very least indicates they

375

are less aware or attuned to local crime. Conversely, LN policy compliance correlated with slightly

376

more development-wide crime, although these latter results were non-significant. Nonetheless,

377

these contrasting findings may help bridge the urban planning and criminology perspectives on

378

the design features that constitute a ‘safe’ neighbourhood. They hint that the design of an

379

‘objectively’ safe place may differ from the design of a ‘subjectively’ safe space. This notion could

380

be further complicated by the nuances of how residents’ perceptions of safety from crime are

381

measured. For example, other research set in these same developments suggests key attributes

382

of a compact walkable neighbourhood (e.g., retail) impact fear of crime (i.e., an emotional

383

reaction to crime) and perceptions of crime (i.e., a cognitive assessment of crime) differently

384

(Foster et al., 2010; Foster et al., 2013a; Foster et al., 2013b). While residents may perceive crime

385

to be present (or even problematic) in their local area, if the crime they perceive does not cause

386

them to feel unsafe, or fearful, it may not negatively impact their day-to-day experiences (Foster &

387

Giles-Corti, 2008).

388

TE D

M AN U

SC

RI PT

374

Limitations:

390

This study had several limitations. First, we focused primarily on self-reported victimisation, which

391

combined three types of crime (i.e., household burglary, harassment or threatening behaviour,

392

physical attack or mugging) and included both direct and indirect victimisation (i.e., hearing about

393

a crime second hand from a friend or family member) (Austin et al., 2002; Hale, 1996). While it

394

could be anticipated that this outcome would better capture local crime than other self-report

395

measures (such as fear of crime or perceptions of crime), it should not be regarded as a proxy for

396

reported crime within the development as: (1) it includes offences that would often not be

397

reported to police (e.g., minor harassment or threatening behaviour); (2) participants who are

398

more fearful about crime could have a heightened awareness and recollection of offences

AC C

EP

389

16

ACCEPTED MANUSCRIPT

committed against both themselves and their friends and family; and (3) it focuses on offences

400

within the neighbourhood, however the area that residents perceive as constituting their

401

‘neighbourhood’ does not necessarily match their ‘development’ (i.e., the scale of analysis for the

402

reported crime outcome). Arbitrarily defined neighbourhoods may bear little resemblance to

403

what a person actually perceives their neighbourhood to be (Chaix et al., 2009; Smith et al., 2010).

404

Nonetheless, the self-report victimisation measure enabled us to examine participants’ relatively

405

recent experiences of crime in the vicinity of their residential development. Second, our sample

406

comprised participants that remained in their new neighbourhoods after 36 months, however we

407

cannot discount the possibility that participants who found crime to be problematic in their new

408

neighbourhoods may have moved away. Third, our study participants were all homeowners living

409

in single family detached housing, typically located in new suburban green field developments

410

with relatively low crime rates. This sample was not representative of the wider Perth

411

metropolitan area as it reflected a group that was both willing and financially able to move to a

412

new home on the urban fringe. However, while it is possible our findings are specific to this

413

somewhat ‘middleclass’ sample, they are nonetheless directly applicable to many new suburban

414

areas unfolding throughout Australia and the United States (US) that typically incorporate

415

conventional design principles. Finally, our study was cross-sectional so causality cannot be

416

determined.

SC

M AN U

TE D

EP

AC C

417

RI PT

399

418

However, this study also has several strengths. Precise, policy-specific measures of new

419

developments were created in GIS – providing objective measures of the on-ground

420

implementation of the LN policy. While overall levels of policy compliance were low (Hooper et

421

al., 2014), our findings confirm that with more faithful adherence to the LN policy requirements,

422

residents tend to experience less crime. Furthermore, our evaluation was comprehensive in its

17

ACCEPTED MANUSCRIPT

423

examination policy compliance and both residents’ reports of victimisation and crimes reported to

424

police.

425

Conclusion

427

This study provides empirical evidence that new urbanist design can impact neighbourhood crime.

428

Our findings revealed that increased compliance with a new urbanist inspired planning policy

429

correlated with lower odds of self-report victimisation among residents. Furthermore, our

430

unpacking of the individual requirements that comprised the policy ‘elements’ underscored the

431

importance of the town centre configuration, a comprehensive network of footpaths (ideally with

432

tree coverage), proximate access to parks, and the provision of diverse housing stock. However, it

433

should also be noted that our findings for policy compliance and actual development-wide crime

434

were positive but non-significant. To some extent, these results bridge the opposing perspectives

435

from planning and criminology disciplines on what constitutes a ‘safe’ neighbourhood. Indeed,

436

this dichotomy raises important questions about the type of environments in which people enjoy

437

and want to live. For some, crime might be a necessary and acceptable trade-off for living in a

438

(potentially) more vibrant, liveable walkable community (Foster et al., 2014).

SC

M AN U

TE D

EP

440

AC C

439

RI PT

426

18

ACCEPTED MANUSCRIPT

References

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491

Austin, D.M., Furr, L.A., & Spine, M. (2002). The effects of neighborhood conditions on perceptions of safety. Journal of Criminal Justice, 30, 417-427. Beavon, D.J.K., Brantingham, P.L., & Brantingham, P.J. (1994). The influence of street networks on the patterning of property offences. In R.V. Clarke (Ed.), Crime Preventions Studies. New York: Criminal Justice Press. Bedimo-Rung, A.L., Mowen, A.J., & Cohen, D.A. (2005). The significance of parks to physical activity and public health: a conceptual model. American Journal of Preventive Medicine, 28, 159-168. Bowes, D.R. (2007). A two-stage model of the simultaneous relationship between retail development and crime. Economic Development Quarterly, 21, 79-90. Brantingham, P.L., & Brantingham, P.J. (1993). Nodes, paths and edges: considerations on the complexity of crime and the physical environment. Journal of Environmental Psychology, 13, 3-28. Burden, D. (2006). 22 Benefits of Urban Street Trees. Glatting Jackson and Walkable Communites, Inc. Cervero, R., & Kockelman, K. (1997). Travel demand and the 3Ds: Density, diversity, and design. Transportation Research Part D: Transport and Environment, 2, 199-219. Chaix, B., Merlo, J., Evans, D., Leal, C., & Havard, S. (2009). Neighbourhoods in eco-epidemiologic research: Delimiting personal exposure areas. A response to Riva, Gauvin, Apparicio and Brodeur. Social Science and Medicine, 69, 1306-1310. Chiesura, A. (2004). The role of urban parks for the sustainable city. Landscape and Urban Planning, 68, 129-138. Christian, H., Knuiman, M., Bull, F., Timperio, A., Foster, S., Divitini, M., et al. (2013). A new urban planning code's impact on walking: The residential environments project. American Journal of Public Health, 103, 1219-1228. City of Melbourne. (2011). Urban Forest Strategy: Making a Greener City 2012-2032. Melbourne, Australia. Congress for the New Urbanism. (2001). Charter of the New Urbanism. Congress for the New Urbanism. (2007). Charter Awards Winners 2001 - 2006. Congress for the New Urbanism. Cozens, P. (2008). New urbanism, crime and the suburbs: A review of the evidence. Urban Policy and Research, 26, 429-444. Cozens, P., & Hillier, D. (2008). The Shape of Things to Come: New Urbanism, the Grid and the Cul-De-Sac. International Planning Studies, 13, 51-73. Cozens, P., & Hillier, D. (2012). Revisiting Jane Jacob's 'eyes on the street' for the twenty-first century: evidence from environmental criminology. In S. Hirst, & D. Zahm (Eds.), The urban wisdom of Jane Jacobs. Oxfordshire, UK: Routledge. Dill, J. (2003). Measuring Network Connectivity for Bicycling and Walking. Joint Congress of ACSP-AESOP. Leuven, Belgium. Dill, J. (2006). Evaluating a New Urbanist Neighborhood. Berkeley Planning Journal, 19. Donovan, G.H., & Prestemon, J.P. (2012). The Effect of Trees on Crime in Portland, Oregon. Environment and Behavior, 44, 3-30. Duany, A., Plater-Zyberk, E., & Speck, J. (2000). Suburban Nation: The Rise of Sprawl and the Decline of the American Dream. New York: North Point Press. Falconer, R. (2008). Transport and Sustainability Study. School of Population Health. Perth: University of Western Australia. Foster, S., & Giles-Corti, B. (2008). The built environment, neighborhood crime and constrained physical activity: an exploration of inconsistent findings. Preventive Medicine, 47, 241-251. Foster, S., Giles-Corti, B., & Knuiman, M. (2010). Neighbourhood design and fear of crime: A socialecological examination of the correlates of residents' fear in new suburban housing developments. Health and Place, 16, 1156-1165. Foster, S., Giles-Corti, B., & Knuiman, M. (2011). Creating safe walkable streetscapes: does house design and upkeep discourage incivilities in suburban neighbourhoods? Journal of Environmental Psychology, 31, 79-88.

AC C

EP

TE D

M AN U

SC

RI PT

441

19

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Foster, S., Knuiman, M., Villanueva, K., Wood, L., Christian, H., & Giles-Corti, B. (2014). Does walkable neighbourhood design influence the association between objective crime and walking? International Journal of Behavioral Nutrition and Physical Activity, 11, 100. Foster, S., Wood, L., Christian, H., Knuiman, M., & Giles-Corti, B. (2013a). Planning safer suburbs: Do changes in the built environment influence residents’ perceptions of crime? Social Science and Medicine, 97, 87-94. Foster, S., Wood, L., Knuiman, M., & Giles-Corti, B. (2013b). Suburban neighbourhood design: associations with fear of crime versus perceived crime. Journal of Environmental Psychology, 36, 112-117. Gorman, D.M., Speer, P.W., Gruenewald, P.J., & Labouvie, E.W. (2001). Spatial dynamics of alcohol availability, neighborhood structure and violent crime. Journal of Studies on Alcohol, 62, 628-636. Gruenewald, P.J., Freisthler, B., Remer, L., LaScala, E.A., & Treno, A. (2006). Ecological models of alcohol outlets and violent assaults: crime potentials and geospatial analysis. Addiction, 101, 666-677. Hale, C. (1996). Fear of Crime: a review of the literature. International Review of Victimology, 4, 79-150. Hooper, P., Giles-Corti, B., & Knuiman, M. (2014). Evaluating the implementation and active living impacts of a state government planning policy designed to create walkable neighborhoods in Perth, Western Australia. American Journal of Health Promotion, 28, S5-S18. Jacobs, J. (1961). The Death and Life of Great American Cities. London: Jonathon Cape. Kim, J., & Kaplan, R. (2004). Physical and psychological factors in sense of community: New urbanist Kentlands and nearly Orchard Village. Environment and Behaviour, 36, 313-340. Knowles, P. (2003). The cost of policing New Urbanism. Community Safety Journal, 2, 33-33+. Kuo, F.E., Bacaicoa, M., & Sullivan, W.C. (1998). Transforming inner-city landscapes: trees, sense of safety, and preference. Environment and Behavior, 30, 28-59. Kuo, F.E., & Sullivan, W.C. (2001a). Agression and violence in the inner city: Effects of environment via mental fatigue. Environment and Behavior, 33, 543-571. Kuo, F.E., & Sullivan, W.C. (2001b). Environment and crime in the inner city. Environment and Behavior, 33, 343-367. Leyden, K.M. (2003). Social capital and the built environment: the importance of walkable neighbourhoods. American Journal of Public Health, 93, 1546-1551. Lund, H. (2003). Testing the claims of new urbanism: local access, pedestrian travel, and neighboring behaviors. Journal of the American Planning Association, 69, 414-429. Maas, J., Spreeuwenberg, P., Van Winsum-Westra, M., Verheij, R.A., de Vries, S., & Groenewegen, P.P. (2009). Is green space in the living environment associated with people's feelings of social safety? Environment and Planning A, 41, 1763-1777. Maruthaveeran, S., & van den Bosch, C.C.K. (2014). A socio-ecological exploration of fear of crime in urban green spaces – A systematic review. Urban Forestry & Urban Greening, 13, 1-18. McCord, E.S., Ratcliffe, J.H., Garcia, R.M., & Taylor, R.B. (2007). Nonresidential crime attractors and generators elevate perceived neighborhood crime and incivilities. Journal of Research in Crime and Delinquency, 44, 295-320. Nasar, J. (1982). A model relating visual attributes in the residential environment to fear of crime. Journal of Environmental Systems, 11, 247-255. Nasar, J., Fisher, B., & Grannis, M. (1993). Proximate physical cues to fear of crime. Landscape and Urban Planning, 26, 161-178. Nasar, J., & Jones, K. (1997). Landscapes of fear and stress. Environment and Behavior, 29, 291-323. Pendola, R., & Gen, S. (2008). Does “Main Street” Promote Sense of Community? A Comparison of San Francisco Neighborhoods. Environment and Behavior, 40, 545-574. Popova, S., Giesbrecht, N., Bekmuradov, D., & Patra, J. (2009). Hours and Days of Sale and Density of Alcohol Outlets: Impacts on Alcohol Consumption and Damage: A Systematic Review. Alcohol and Alcoholism, 44, 500-516. Rodríguez, D.A., Khattak, A.J., & Evenson, K.R. (2006). Can New Urbanism Encourage Physical Activity? American Planning Association. Journal of the American Planning Association, 72, 43. Sallis, J.F., Floyd, M.F., Rodríguez, D.A., & Saelens, B.E. (2012). Role of Built Environments in Physical Activity, Obesity, and Cardiovascular Disease. Circulation, 125, 729-737. Schneider, R.H., & Kitchen, T. (2007). Crime Prevention and the Built Environment. New York: Routledge.

AC C

492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544

20

ACCEPTED MANUSCRIPT

SC

RI PT

Smith, G., Gidlow, C., Davey, R., & Foster, C. (2010). What is my walking neighbourhood? A pilot study of English adults' definitions of their local walking neighbourhoods. International Journal of Behavioral Nutrition and Physical Activity, 7, 34. Song, Y. (2005). Smart Growth and Urban Development Pattern: A Comparative Study. International Regional Science Review, 28, 239-265. Talen, E. (2002). The social goals of new urbanism. Housing Policy Debate, 13, 165-188. Talen, E., & Koschinsky, J. (2014). Compact, Walkable, Diverse Neighborhoods:Assessing Effects on Residents. Housing Policy Debate, 24, 717-750. Western Australian Planning Commission. (2000). Liveable Neighbourhoods. Perth: State of Western Australia. Western Australian Planning Commission. (2009). Street Trees and Utility Planning Discussion Paper. Perth, WA. Wolch, J.R., Byrne, J., & Newell, J.P. (2014). Urban green space, public health, and environmental justice: The challenge of making cities ‘just green enough’. Landscape and Urban Planning, 125, 234-244. Wood, L., Frank, L.D., & Giles-Corti, B. (2010). Sense of community and its relationship with walking and neighborhood design. Social Science and Medicine. 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 built environment, social capital and residents' perceptions of safety. Health and Place, 14, 15-31. Zelinka, A., & Brennan, D. (2001). Safescape. Chicago: American Planning Association.

M AN U

545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565

AC C

EP

TE D

566

21

ACCEPTED MANUSCRIPT

Table 1 Characteristics of the 603 RESIDE participants from the 36 residential developments Characteristic

Gender (male) Mean age (SD) Education Secondary or less Trade/certificate Bachelor or higher Marital status Married/defacto Separated/divorced Single Children at home (yes) Victimisation (yes)

Percentage 37.5% 43.2 (11.7) 33.3% 39.3% 27.4%

RI PT

567

86.7% 9.6% 3.6% 52.6% 28.9%

SC

568

AC C

EP

TE D

M AN U

569

22

Victimisation1 OR (95% CI)

LN compliance

Mean (SD)

Total LN compliance

47.23 (7.07)

0.60 (0.53,0.67)

0.000

Community Design

37.21 (22.22)

0.94 (0.83,1.05)

0.264

Movement Network

46.96 (4.79)

0.49 (0.31,0.76)

0.001

Lot Layout

45.99 (18.35)

0.84 (0.77,0.93)

Public Parkland

50.82 (11.07)

0.78 (0.67,0.90)

RI PT

p

0.000 0.001

TE D

M AN U

SC

*Separate models depict the increase in odds of victimisation for each 10% increase in the level of policy compliance. 1 All models control for age, gender, education, marital status, number of children, area socio-economic status, stage of construction and size of development. Bold denotes p≤0.05.

EP

571 572 573 574

ACCEPTED MANUSCRIPT

Table 2: Associations between LN policy compliance, overall and by element, and victimisation.

AC C

570

23

575 576

ACCEPTED MANUSCRIPT

Table 3: Associations between compliance with specific design requirements from each LN element associated and victimisation. Victimisation Liveable Neighbourhoods Policy Requirement n

OR (95% CI)

p

Community Design None

93

1.00

Big Box

395

1.42 (1.03, 1.96)

0.031

Main St

115

0.93 (0.65, 1.32)

0.676

603

0.98 (0.96, 1.00)

0.017

603

0.37 (0.17, 0.78)

0.010

Movement Network 1

Sidewalk to road ratio [OR for 1 unit increase in sidewalk:road ratio] 2

SC

Tree density along footpaths [OR for 1 unit increase in number of trees per km of footpath] Lot Layout 3

Residential dwelling density (development wide) [OR for 1 unit increase in #dwellings ÷ land (ha) zoned residential]

Mean residential lot size [OR for 1 unit increase in mean lot size] 4

603

Number of different lot sizes present [OR for 1 unit increase in different lot sizes present]

TE D

Public Parkland

1.24 (1.16, 1.32)

M AN U

2

Percent of residential lots ≤350m [OR for 1 unit increase]

RI PT

Configuration of neighbourhood centre ≤1600m

Percent houses within 400m of any park [OR for 1 unit increase]

0.000

603

0.97 (0.95, 1.00)

0.031

603

1.01 (1.01, 1.01)

0.000

603

0.72 (0.56, 0.95)

0.018

603

0.98 (0.97, 0.99)

0.000

603

0.99 (0.99, 1.00)

0.043

603

0.99 (0.99, 1.00)

0.000

5

Percent houses within 400m of a local neighbourhood park [OR for 1 unit increase]

Final models for each Liveable Neighbourhoods element derived from backwards stepwise elimination. Models depict the increase in odds of victimisation for each 10% increase in the level of compliance with each specific requirement. All models adjust for demographic characteristics (age; gender; education, marital status, children ≤18 years and under living at home), area socio-economic status, stage of construction and scale of development. Bold denotes p≤0.05. 1 Sidewalk to road ratio = length of all footpath segments alongside/adjacent to/parallel with the roads ÷ length of all roads. 2 Tree density along footpaths = number of trees along footpaths (within a 5m buffer) ÷ length (km) of footpaths within the development. 3 Residential density (development wide) = Net residential dwelling density = number of residential dwellings ÷ area (ha) of residentially zoned (and constructed) land. 4 Lot mix score = the number of different lot sizes present (categories: ≤350m2; >350 - ≤550m2; >550 - ≤750m2; >750 ≤950m2; >950 m2) (value from 1-5). 5 Local neighbourhood park is classified as 0.3 ≤ 0.5 ha. 6 A large neighbourhood park is classified as 1.5 ≤ 2.5 ha. Note: ‘Sidewalk’ refers to footpaths adjacent to roads, whereas ‘footpath’ refers to all sealed paths regardless of location.

AC C

577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595

6

EP

Percent houses within 400m of a large neighbourhood park [OR for 1 unit increase]

24

597

ACCEPTED MANUSCRIPT

Table 4: Associations between LN policy compliance, overall and by element, and crime reported within the development.

LN compliance

Crimes against person in public RR CI p

Total LN compliance

1.39

0.76,2.54

0.281

1.58

0.71,3.50

0.259

Community Design

1.05

0.83,1.32

0.692

1.11

0.86,1.47

0.420

Movement Network

0.93

0.49,1.76

0.814

1.55

Lot Layout

1.24

0.96,1.60

0.107

1.11

Public Parkland

1.07

0.82,1.40

0.621

1.13

0.374

0.76,1.62

0.573

0.74,1.74

0.565

SC

0.59,4.05

Separate negative binomial log-linear models for number of crimes adjust for area socio-economic status, stage of construction and scale of development, with log of number of dwellings included as an offset term. RR: Relative increase in crime rate per dwelling associated with a 10% increase in compliance. Crime against the person in public space: Range 0-26; mean=3.03 (SD=5.69) Actual and attempted burglary: Range 0-146; mean=23.22 (SD=31.00)

M AN U

598 599 600 601 602 603

Actual & attempted burglary RR CI p

RI PT

596

604 605

AC C

EP

TE D

606

25

ACCEPTED MANUSCRIPT Research Highlights

There is limited evidence that new urbanism design principles can help limit crime.

RI PT

We tested the impact of compliance with a new urban style planning policy on crime. For each 10% increase in compliance, self-report victimisation fell by 40% (p=0.000).

SC

Associations between policy compliance and objective crime were non-significant.

AC C

EP

TE D

M AN U

New urbanism may deliver additional social and wellbeing benefits for local residents.