Available online at www.sciencedirect.com

(2008) 265–269www.elsevier.com/locate/ypmed

Preventive Medicine 47

Travel perceptions, behaviors, and environment by degree of urbanization

Hannah M. Badland a,⁎, Mitch J. Duncan b, W. Kerry Mummery b

a Centre for Physical Activity and Nutrition Research, Auckland University of Technology, Auckland, New Zealandb Centre for Social Science Research, Central Queensland University, Rockhampton, Australia

Available online 9 February 2008

Abstract

Objectives. This study examines how engagement in transport-related physical activity (TPA), perceptions of the commute route, actual traveldistance, and perceptions of travel distance vary by degree of urbanization in rural and remote areas of Central Queensland, Australia.

Method. A random, cross-sectional sample of 1230 adults aged 18 years and over residing in rural and remote Central Queensland weresurveyed by telephone in October–November 2006. Engagement in and attitudes towards TPA, perceptions of the commute route, actual traveldistances, and perceived appropriate travel distances for TPA were assessed.

Results. For this study 765 adults were eligible for the analysis as they traveled to an occupation. Overall, 86% of the sample commuted to theiroccupation by motorized modes, 71% of individuals recognized 5 km or less as an appropriate distance to travel via TPA modes, and 45% ofrespondents reported travel distances less than five kilometers (km) to reach their workplace. Participants from small rural and remote settingsmore frequently traveled greater than 20 km to reach their occupation, reported fewer sidewalks and shops, and less heavy traffic enroutecompared to respondents in large rural settings.

Conclusion. Infrastructure for, and participation in TPA varies according to degree of urbanization. Future work in this area should target thesubstantial population that recognizes they could engage in TPA. As such, increasing TPA participation in these settings likely requires a culturalshift and infrastructure improvements.© 2008 Elsevier Inc. All rights reserved.

Keywords: Perceived travel distance; Urbanization; Transport-related physical activity; Infrastructure

Introduction

Increasing population-level physical activity engagement isan important public health priority for many developedcountries (Department of Health Physical Activity HealthImprovement and Promotion, 2004; US Department of Healthand Human Services, 2000), yet many countries continue toreport low levels of activity engagement (US Department ofHealth and Human Services, 2000; World Health Organization,2004). Understanding the contextual environment and provid-ing appropriate infrastructure that supports physical activity areimportant steps to addressing these current trends (Saelens et al.,2003; Badland and Schofield, 2005).

Research has shown that distinct aspects of the physicalenvironment facilitate or inhibit different types of physical

⁎ Corresponding author. Fax: +64 9 921 9746.E-mail address: hannah.badland@aut.ac.nz (H.M. Badland).

0091-7435/$ - see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.ypmed.2008.01.021

activity (Lee and Moudon, 2006; Pikora et al., 2006). AnAustralian study identified that destination accessibility (shops,public transport stops) was positively associated with self-reported walking for transport in adults, whereas a high level ofenvironmental functionality (paths maintained, surfaced, con-tinuous, and direct) was positively related with recreationalwalking (Pikora et al., 2006). Similarly, Lee and Moudon(2006) compared objectively-measured (determined usinggeographical information systems (GIS)) environmental vari-ables with walking in US adults. Walking for transport waspositively associated with presence of utilitarian destinations,and negatively associated with presence of hills. Conversely,recreational walking was positively related to hill and sidewalkpresence. Another important consideration when examining theassociation between the environment and physical activity,particularly transport-related physical activity (TPA) engage-ment, is commute distance. Recent work has shown occupa-tional commute distance was the most important contributor to

266 H.M. Badland et al. / Preventive Medicine 47 (2008) 265–269

work-related TPA engagement (Cerin et al., 2007), anddistances less than 5 km were generally perceived as acceptablefor TPA engagement amongst adults (Badland et al., 2007).

Overall physical activity levels and perceived environmentalbarriers to activity engagement have been compared by townsize (Badland and Schofield, 2006). Although town size showedmixed associations with total physical activity accumulation,clear relationships were identified with barriers to physicalactivity engagement by region. Lack of infrastructure wasreported as a significant barrier to overall physical activity en-gagement for residents of small towns, whereas time constraintswere more frequently reported for residents from large cities(Badland and Schofield, 2006). Other studies have shown a lowerproportion of rural residents reported the presence of micro-scale infrastructure such as sidewalks, recreation facilities, andparks, and were less active overall when compared to theirmetropolitan counterparts (Wilcox et al., 2000; Parks et al.,2003).

In summary, the aforementioned studies identified disparatephysical activity domains are associated with different environ-mental characteristics, and regional settings exhibit diverseenvironments that can facilitate or inhibit physical activity en-gagement. This study seeks to build upon the existing literatureby specifically investigating occupation-related TPA behaviorsand perceptions by degree of urbanization for large and smallrural settings and remote areas, commute distance, and per-ceptions of physical environment variables enroute for CentralQueensland adults.

Methods

The Central Queensland Social Survey (CQSS) is an annual omnibuscomputer-assisted-telephone-interview (CATI) survey assessing the health ofadults (N18 years of age) residing in Central Queensland, Australia, and isconducted by the Population Research Laboratory, Central QueenslandUniversity. Participants (n=1230) were randomly selected with replacementfrom the electronic white pages; i.e., if a participant could not be contacted orrefused to participate, another telephone number was randomly generated. Due toomnibus nature of the survey, the CQSS was powered to provide an accuratesample of the overall survey area to within 2.8% points 19 times out of 20. It wasnot designed to provide a representative sample of TPA behaviors acrossgeographical locations. The survey was conducted in October 2006. Theresponse rate was 41%, which was comparable to other recently conducted CATIsurveys (Caperchione et al., in press; Mummery et al., 2007; Ainsworth et al.,2003; Badland and Schofield, in press). The Central Queensland UniversityHuman Research Ethics Committee approved the study and all participantsprovided verbal consent.

The CQSS 2006 assessed leisure time physical activity, TPA frequency,perceptions of micro-scale characteristics of the travel route and at place ofwork/study, TPA attitudes, perceived distances appropriate for TPA, and socio-demographic characteristics. Those who did not work (n=431) or hadincomplete travel information (n=34) were excluded from this analysis, leavinga final sample of n=765.

TPA engagement

The question: “In a usual week, how often would you walk or cycle for themajority of the commute distance from your home to or from your usual place ofwork/study?” was used to assess TPA engagement frequency. Clarification wasprovided that travel to and from the workplace was classified as two separatejourneys. Responses were collapsed into four categories— 0 trips, 1–2 trips, 3–5 trips and N5 trips.

Perceived distance appropriate for TPA engagement

Respondents were asked: “In kilometers, how far do you think it isappropriate for you to walk or cycle to reach a destination if you have the time, itwas pleasant weather, and you did not have to carry anything?” to assessperceptions of distance appropriate for TPA engagement. Responses werecollapsed into three kilometer (km) categories. This question was derived from apreviously developed item (Rafferty et al., 2004) to assess general perceptions ofTPA.

Perceived micro-scale characteristics

Respondents rated the presence of shops/commercial premises, footpaths,and heavy traffic along the commute route using a five-point Likert scale(strongly agree (1) to strongly disagree (5)). Responses were dichotomouslyrecoded into categories: ‘agree’ (strongly agree, agree (1)) and ‘not agree’(neither, disagree, strongly disagree (0)).

Physical activity engagement

Overall physical activity and non-specific destination TPA were assessedusing items based on the Active Australia Questionnaire (Australian Institute ofHealth and Welfare, 2003). This questionnaire is a reliable instrument (Brownet al., 2004) that is used in national surveys to determine physical activity levelsof Australian adults. The questionnaire assessed the frequency and duration oftime spent in recreational and transportation walking, and moderate andvigorous intensity physical activity separately. Participation in sufficient levelsof physical activity was determined using the minimum threshold of 150 minreported for all physical activities over five or more sessions per week(Department of Health, Physical Activity, Health Improvement, and Promotion,2004).

Attitudes towards TPA

Participants rated their agreement with two questions assessing perceivedhealth benefits and enjoyment of TPA using a five-point Likert scale (stronglyagree (1) to strongly disagree (5)). The questions: “I can gain benefits to myhealth by regularly participating in walking/cycling for transport” and “I enjoywalking/cycling for transport purposes” were asked. Responses were dichot-omously recoded into categories: ‘agree’ and ‘not agree’ using the same formatas used for micro-scale characteristics.

Town classifications

The regional, remote, and metropolitan areas (RRMA) classification is usedas a proxy measure of access to classify geographic areas in Australia(Australian Government Department of Health and Ageing, 2005). Areas wereclassified into one of the following seven categories: 1) capital cities, 2) other(non-capital) metropolitan centers (≥100,000 residents), 3) large rural centers(25,000–99,999 residents), 4) small rural centers (10,000–24,999 residents), 5)other rural areas (b10,000 residents), 6) remote centers (≥5000 residents), and7) other remote areas (b5000 residents) (Australian Institute of Health andWelfare, 2004). Respondents' degrees of urbanization were determined bylinking their reported postal code with the associated RRMA classification.RRMA classifications of 3, 4–5, and 6–7 were collapsed to create the ‘largerural’, ‘small rural’, and ‘remote’ classifications, respectively.

Statistical analyses

Comparisons between groups in contingency tables were made usingPearson's χ2 test. Adjusted (controlling for gender, age, education, and annualhousehold income) odds ratios (OR) were derived using multinomial logisticregression analyses to examine relationships between travel and physicalactivity variables for those who lived in small rural or remote areas versus thoseresiding in large rural regions. The independent variable for this analysis was thedegree of urbanization. Percent agreement values were used to determine

Table 1Socio-demographic profile by degree of urbanization of Central Queenslandadults surveyed in October–November 2006

Town classification

Variable n (%) Largerural

Smallrural

Remote p-value

n (%) n (%) n (%)

Gender 0.57Male 409 (53) 219 (52) 140 (55) 49 (54)Female 356 (47) 201 (48) 114 (45) 42 (46)

Age (years) 0.8618–34 159 (21) 100 (24) 38 (15) 21 (23)35–44 226 (30) 103 (25) 91 (36) 31 (34)45–54 235 (31) 143 (34) 67 (27) 25 (28)≥55 144 (19) 74 (18) 57 (23) 13 (14)

Years of education 0.26≤10 208 (27) 98 (24) 81 (32) 29 (32)11–12 185 (24) 107 (26) 61 (24) 16 (18)13–14 113 (15) 69 (17) 33 (13) 11 (12)≥15 251 (33) 140 (34) 78 (31) 34 (38)

Annual household income (AUD$) ≤0.001b26,000 64 (11) 38 (12) 21 (12) 4 (6)26,001–52,000 106 (19) 73 (23) 28 (16) 6 (9)52,001–100,000 226 (40) 139 (43) 70 (39) 17 (25)N100,000 171 (30) 70 (22) 59 (33) 42 (61)

267H.M. Badland et al. / Preventive Medicine 47 (2008) 265–269

differences in environmental perceptions across the three regional categories.Logistic regression and percent agreement analyses were conducted for theoverall sample and for those who reported commuting 5 km or fewer to theiroccupation; however, only those data for the entire sample are presented in thetables. Statistical comparisons were made using SPSS version 14.0. An α=0.05was used to define statistical significances.

Table 2Adjusted logistic regression analysis with TPA variables by degree of urbanization o

Town classification

Variable n (%) Large rural

(%) OR a (95%CI)

Non-specific TPA engagement (mins/week)b30 560 (73) (72) 1.00 ref30–149 135 (18) (20) 1.00 ref≥150 70 (9) (8) 1.00 ref

Non-specific appropriate TPA distance (km)b2 102 (13) (13) 1.00 ref2–5 441 (58) (57) 1.00 refN5 222 (29) (30) 1.00 ref

Work-related TPA engagement (sessions/week)0 655 (86) (86) 1.00 ref1–2 36 (5) (6) 1.00 ref3–5 35 (5) (4) 1.00 refN5 37 (5) (5) 1.00 ref

Work-related commute distance (km)≤5 331 (45) (45) 1.00 ref6–10 139 (19) (26) 1.00 ref11–20 120 (16) (17) 1.00 refN20 153 (21) (12) 1.00 ref

Overall physical activity levelsSedentary 126 (17) (18) 1.00 refInsufficiently active 273 (36) (34) 1.00 refSufficiently active 364 (48) (49) 1.00 ref

aAdjusted by gender, age, education, annual household income.Key: CI=confidence interval; km=kilometer; OR=odds ratio.

Results

Overall, the sample (n=765) consisted of slightly more males(53%) than females; the majority of respondents were between35–54 years of age; 33% of respondents held a tertiaryqualification; and the modal household income was $52,001–$100,000 Australian dollars per annum (Table 1). Respondentswere further separated by degree of urbanization into threecategories: large rural (55%; n=420), small rural (33%; n=254),and remote (12%; n=91) center residents. Significant differencesexisted for household income when degree of urbanization wasexamined, with those living in remote settings being more likelyto report higher earnings.

The majority of respondents reported no work-related TPAsessions and engaged in less than 30 min of non-specificdestination TPA per week, yet perceived 2–5 km as appropriatedistances for TPA engagement. Approximately half of thesample was classified as sufficiently active for health benefits(Table 2). Significant differences across town sizes were shownfor work-related commute distances. Individuals from smallrural and remote settings were more likely to live more than20 km away from their occupation when compared to those fromlarge rural areas. In addition, residents in small rural settingswere more likely to engage in 3–5 sessions of work-related TPAper week than those from large rural settings. This associationremained significant when the analyses were restricted to thosewho traveled 5 km or fewer to access their occupation (OR=3.44; 95% CI=1.29, 9.15). Other significant associations shownin Table 2 were that when compared with participants from large

f Central Queensland adults surveyed in October–November 2006

Small rural Remote

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

(76) 1.33 (0.86–2.06) (74) 0.89 (0.49–1.61)(14) 0.56 (0.33–0.95) (18) 0.94 (0.47–1.89)(11) 1.34 (0.70–2.56) (9) 1.49 (0.62–3.59)

(15) 1.31 (0.78–2.20) (11) 0.74 (0.31–1.79)(58) 1.04 0.71–1.51) (60) 1.24 (0.72–2.15)(27) 0.81 (0.53–1.22) (29) 0.90 (0.49–1.63)

(85) 0.89 (0.54–1.47) (90) 1.81 (0.71–4.61)(3) 0.49 (0.19–1.29) (4) 0.26 (0.04–1.82)(6) 2.32 (1.01–5.34) (3) 0.95 (0.21–4.21)(5) 1.10 (0.49–2.48) (2) 0.74 (0.17–3.14)

(41) 0.86 (0.58–1.26) (52) 1.56 (0.89–2.75)(12) 0.42 (0.25–0.72) (4) 0.05 (0.01–0.32)(20) 1.20 (0.76–1.91) (2) 0.16 (0.04–0.62)(28) 2.37 (1.47–3.81) (42) 4.92 (2.65–9.13)

(14) 0.84 (0.51–1.38) (18) 0.88 (0.43–1.81)(38) 1.18 (0.80–1.74) (41) 1.76 (1.02–3.04)(48) 0.95 (0.67–1.38) (41) 0.62 (0.36–1.08)

Table 3Perceived environmental barriers to TPA by degree of urbanization in CentralQueensland adults surveyed in October–November 2006

Town classification

n (%) Largerural

Smallrural

Remote p-value

Agree%

Agree%

Agree%

No sidewalks along commuteroute

354 (46) 39 59 53 ≤0.001

Poor mixed land use alongthe commute route

253 (33) 23 47 51 ≤0.001

Heavy traffic along thecommute route

563 (74) 83 71 57 ≤0.001

Do not enjoy TPA engagement 237 (31) 31 32 37 0.34Do not perceive health gainsfrom TPA engagement

30 (4) 5 4 2 0.26

268 H.M. Badland et al. / Preventive Medicine 47 (2008) 265–269

rural settings, small rural respondents were less likely to engagein 30–149 min of non-specific TPA per week and respondentsfrom remote areas were more likely to be classed as insuf-ficiently active overall. These somewhat conflicting findings forsmall rural respondents may have existed because of type I errorsas a large number of comparisons were made within the table andthe small sample sizes existed.

Macro- and micro-level infrastructural barriers to TPAengagement differed by degree of urbanization, with a higherproportion of residents in small rural or remote settings beingsignificantly more likely to report lack of sidewalks and poormixed land use along the commute route to their occupation(p-value=≤0.001) (Table 3). Across all groups heavy trafficwas the most common barrier to TPA engagement, with thosewho lived in remote settings being least likely to report heavytraffic along the commute route (p-value=≤0.001). No differ-ences existed by degree of urbanization with perceptionsand enjoyment of TPA. The trends shown for mixed land use(p-value=≤0.01) and heavy traffic (p-value=≤0.001) re-mained when commute distances were limited to 5 km or fewer.

Discussion

This paper describes how participation in occupation-relatedTPA and perceptions of the commute route vary by degree ofurbanization. Unique aspects of the study are the examination ofenvironmental variables by overall commute journey, with a sub-analysis for distances of 5 km or fewer, and the focus on rural andremote settings. Occupation-related TPA engagement remainedlow, regardless of commute distance or geographical location.Those who resided in small rural settings were more likely toengage in 3–5 occupation-related TPA sessions per week whencompared to residents from large rural settings, yet commutedistances greater than 20 km were positively associated withresiding in smaller regions. One potential explanation for thisrelationship may be that the observed associations are linked tothe compactness of town design in small rural areas and also thewider catchment areas that the settings service (Ewing et al.,2003).

Despite overall TPA engagement being low, 2–5 km wasperceived by the majority of the population as appropriatedistances for TPA engagement and 45% of the overall samplereported commuting 5 km or fewer to access their occupation.Based on these figures, if future interventions can harness thissupport for TPA engagement at these distances and convert it tobehavior change, it is likely that substantial public health andenvironmental gains will be made. The majority of the sam-ple also recognized health benefits could be gained by TPAengagement.

The physical environment was compared by region. A higherproportion of those residing in small rural and remote locationsreported the absence of sidewalks and poorer mixed land usealong commute routes when compared to those living in largerural locations. The sidewalk relationship remained significantfor commute distances 5 km or fewer. The association of reducedsidewalk availability in smaller areas is similar to that previouslyreported in US (Wilcox et al., 2000) and New Zealand (Badlandand Schofield, 2006) populations.

Heavy traffic was the most common micro-scale variablecited; yet differences by degree of urbanization were evident.Individuals residing in large rural settings were more likely toreport the presence of heavy traffic along commute routeswhen compared to the other areas. This finding was similar toprevious research that demonstrated residents of large citieswere most likely to report heavy traffic volumes as a barrier tophysical activity participation (Badland and Schofield, 2006).Badland and Schofield (2006) however, also demonstratedthat respondents who lived in small towns (b1000 residents)reported high traffic volume as a physical activity barrier. Thedifferences between reported sidewalk presence and degree ofurbanization in the present study, combined with high trafficvolume as a barrier to TPA participation (Dellinger and Staunton,2002) reinforces the need to provide appropriate infrastructure innational strategies to increase TPA engagement (Austroads Inc.,2005).

Study limitations

This study revealed a number of key associations betweendegree of urbanization, perceptions of and occupation-relatedTPA engagement, and urban design variables along the commuteroute. Although useful recommendations have been made basedon these findings, limitations exist. These included the relianceon self-report measures, the cross-sectional study design couldnot determine causality and a TPA measure was not able todistinguish between walking and cycling. The measure of per-ceived travel distance appropriate for TPA engagement was notdestination specific, and potentially much of the sampleconsidered only walking when referring to distances appropriatefor commuting by TPA. A lack of variability was shown fortravel behaviors, although this a reflection of the low level ofengagement observed, and the assessment of TPA may have ledto an under-representation of occupation-related TPA engage-ment. Furthermore, other possible barriers to TPA were notassessed, and other potential confounders such as health statusand body size were not controlled for in the analyses. Future

269H.M. Badland et al. / Preventive Medicine 47 (2008) 265–269

research should seek to incorporate objective measures wherepossible (such as GIS and global positioning systems), and toover-sample, or conduct extensive subgroup analysis, for re-spondents who travel by TPA modes.

Conclusions

The comparisons between degree of urbanization by TPAengagement and built environment infrastructure along the com-mute route for overall travel and for distances restricted to 5 km orfewer revealed useful findings. Disparities existed between lowerlevels of occupation-related TPA engagement and higherperceptions of appropriateness of the behavior by degree ofurbanization. Progressing perceptions of TPA engagement tobehavior change for shorter commute distances will likely resultin substantial population health and environmental gains. Further-more, physical environmental barriers to TPA engagement variedby degree of urbanization, with those residing in large ruralsettings reportingmore problemswith traffic volume, and those insmaller settings reporting lack of sidewalks and poorermixed landuse as inhibiting factors. As such, these findings suggest inter-ventions to increase TPA engagement should focus on alteringsocietal norms towards TPA engagement and manipulating thephysical environment on a regionally specific basis. These con-clusions drawn are similar to previously described approachesoutlining the importance of both environmental and personalcorrelates of activity (Ball et al., 2006; Giles-Corti and Donovan,2003).

Acknowledgments

This research was supported by the Population ResearchGrant Scheme 2006, Centre for Social Science Research,Central Queensland University. HB is supported by a NewZealand National Heart Foundation Research Fellowship.

References

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

Australian Government Department of Health and Ageing, 2005. Review of therural, remote, and metropolitan areas (RRMA) classification: Discussionpaper. Australian Government Department of Health and Ageing, Canberra.

Australian Institute of Health and Welfare, 2003. The Active Australia Survey:A guide and manual for implementation, analysis and reporting. AustralianInstitute of Health and Welfare, Canberra.

Australian Institute of Health and Welfare, 2004. Rural, regional, and remotehealth: A guide to remoteness classifications. Australian Institute of Healthand Welfare, Canberra.

Austroads Inc., 2005. The Australian national cycling strategy 2005–2010.Austroads Incorporated, Sydney.

Badland,H., Schofield, G., 2005. Transport, urban design, and physical activity: anevidence based-update. Transp. Res., Part D Transp. Environ. 10, 177–196.

Badland, H., Schofield, G., 2006. Understanding the relationship between town sizeand physical activity levels: a population study. Health and Place 12, 538–546.

Badland, H., Schofield, G., in press. Health associations with transport-relatedphysical activity and motorized travel to destinations. International Journalof Sustainable Transport.

Badland, H., Schofield, G., Schluter, P., 2007. Objectively measured commutedistance: associationswith actual travelmodes and perceptions to place ofworkor study in Auckland, New Zealand. J. Phys. Activity and Health 4, 80–86.

Ball, K., Timperio, A., Crawford, D., 2006. Understanding environmental influenceson nutrition and physical activity behaviors: where should we look and whatshouldwe count? Int. J. Behav.Nutrition andPhys.Activity 3. doi:10.2286/1479.

Brown, W., Trost, S., Bauman, A., Mummery, W., Owen, N., 2004. Test-retestreliability of four physical activity measures used in population surveys.J. Sci. Med. Sport 7, 205–215.

Caperchione, C., Lauder, W., Kolt, G., Duncan, M., Mummery, W., in press.Associations between social capital and health status in an Australianpopulation. Psychol. Health Med.

Cerin, E., Leslie, E., du Toit, L., Owen, N., Frank, L., 2007. Destinationsthat matter: associations with walking for transport. Health & Place 13,713–724.

Dellinger, A.M., Staunton, C.E., 2002. Barriers to children walking and biking toschool — United States 1999. J. Am. Med. Assoc. 288, 1343–1344.

Department of Health Physical Activity Health Improvement and Promotion,2004. At least five a week: evidence on the impact of physical activity and itsrelationship to health. A report from the Chief Medical Officer. Departmentof Health, London.

Ewing, R., Schmid, T., Killingsworth, R., Zlot, A., Raudenbush, S., 2003.Relationship between urban sprawl and physical activity, obesity andmorbidity. Am. J. Health Promot. 18, 47–57.

Giles-Corti, B., Donovan, R., 2003. Relative influences of individual, socialenvironmental, and physical environmental correlates of walking. Am. J.Public Health 93, 1583–1589.

Lee, C., Moudon, A., 2006. Correlates of walking for transportation orrecreation purposes. J. Phys. Activity and Health 3, S77–S98.

Mummery, W., Duncan, M., Kift, R., 2007. Socio-economic differences in publicopinion regarding fluoridation in Queensland. Aust. N. Z. J. Public Health 31,336–339.

Parks, S.E., Housemann, R.A., Brownson, R.C., 2003. Differential correlatesof physical activity in urban and rural adults of various socioeconomicbackgrounds in the United States. J. Epidemiol. Community Health 57,29–35.

Pikora, T., Giles-Corti, B., Knuiman, M., Bull, F., Jamrozik, K., Donovan, R.,2006. Neighborhood environmental factors correlated with walking nearhome: using SPACES. Med. Sci. Sports Exerc. 38, 708–718.

Rafferty, A.P., McGee, H.B., Petersmarck, K.A., Miller, C.E., 2004. Proportion oftrips made by walking: estimating a state-level baseline for HealthyPeople 2010 objective 22-14. Am. J. Health Promot. 18, 387–391.

Saelens, B.E., Sallis, J.F., Frank, L.D., 2003. Environmental correlates ofwalking and cycling: findings from the transportation, urban design, andplanning literatures. Annals Behav. Med. 25, 80–91.

US Department of Health and Human Services, 2000. Healthy People 2010:Understanding and improving health. US Department of Health and HumanServices, Washington DC.

Wilcox, S., Castro, C., King, A.C., Housemann, R., Brownson, R.C., 2000.Determinants of leisure time physical activity in rural compared with urbanolder and ethnically diverse women in United States. J. Epidemiol. Communityhealth 54, 667–672.

World Health Organization, 2004. Global strategy on diet, physical activity andhealth. World Health Organization, Geneva.