Neighborhood Perceptions and Active Commuting to School Among ...

Journal of Physical Activity and Health, 2010, 7, 257-266 © 2010 Human Kinetics, Inc.

Neighborhood Perceptions and Active Commuting to School Among Adolescent Boys and Girls Norah M. Nelson and Catherine B. Woods Background: Active commuting (AC) to school can increase daily minutes of physical activity yet research is lacking on its determinants. This study examined perceptions of the physical environment as a correlate of AC among adolescents. Methods: Cross-sectional data were collected from 1143 males and 1016 females (mean age 16.04 ± 0.66) who lived within 2.5 miles of their school. Participants’ self-reported active (walk or cycle) or inactive (car, bus, or train) mode of travel to school and perceptions of their neighborhood environment. Bivariate logistic regression examined perceived environmental features associated with active versus inactive modes, adjusted for sociodemographic factors. Significant variables were examined in multivariate models, adjusted for population density and distance. Results: Positive correlates of AC included well-lit streets, landuse-mix diversity, access to shops/public transport, the presence of public parks/bike lanes, and accessible well-maintained paths. Connectivity was unrelated to mode choice. In multivariate analyses, land-use-mix diversity, and the perceived presence of public parks remained significant among males, whereas excess traffic speed, shops within walking distance, and paths separate from the road remained significant among females. Conclusions: Environmental characteristics were associated with active commuting to school, however research must address methodological issues before making recommendations for intervention. Keywords: environment, active transport, youth, gender Regular physical activity is associated with physical, social, and psychological health among young people,1 and the promotion of physical activity across the lifespan is a public health priority.2 Despite this, research shows an age related decline in physical activity among youth3,4 which is typically more pronounced among adolescent girls.5 Active travel (ie, walking or cycling to school) is a promising area for intervention as walking and cycling are feasible and dependable activities through which many young people, including sedentary or irregularly active individuals, can increase daily minutes of physical activity.6–9 Recent research indicates that adolescents who walk10 and cycle11 to school also have higher aerobic capacity than those who travel by motorized transport. For effective promotion of active commuting to school, its determinants must be identified, particularly factors that are amenable to change.12 Social-ecological theory proposes that physical activity is determined by the complex multilevel interaction of multiple demographic, physical, psychological, behavioral, social, and physical environmental factors.13–15 A basic principle Nelson is with the Dept of Sport, Culture, and the Arts, University of Strathclyde, Glasgow, Scotland, United Kingdom. Woods is with the School of Health and Human Performance, Dublin City University, Dublin, Ireland.

underlying this theory is that physical environments influence behavior. Two review papers document environmental correlates of physical activity among children and adolescents3,16 However, separate measurement of physical activity by purpose (ie, for transport, recreation or sport) is important.17 People engage in travel to participate in activities at destinations, whereas people engage in leisure time physical activity for its own sake, therefore elements of the built environment are likely to affect the two differently.17,18 A recent review identified 24 studies that specifically examined environmental correlates of active travel among youth aged 5 to 18 years.19 Panter and colleagues (2008) propose a conceptual framework that integrates the environment into a wider decision making process around travel choices for children and adolescents. Both age and gender are identified as main moderators that may alter the strength of associations between the environment and active travel. Although many studies have examined age related differences in effects,20–23 fewer studies have examined gender differences,17 and this is highlighted as an area warranting further research. Therefore, this study samples males and females in late adolescence (15 to 17 years), a period with potential for increased independent mobility but noted as a period of declining physical activity, particularly among girls. With rates of active commuting to school steadily decreasing,24–26 research on environmental characteristics can inform interventions to reduce and remove

257

258 Nelson and Woods

key environmental barriers, making active commuting easier and more realistic. The purpose of this study was to (i) examine the associations between perceptions of the neighborhood environment and active commuting to school among male and female adolescents, and (ii) identify gender specific features positively or negatively associated with active commuting to inform future research and intervention.

distance (miles) of the journey and distance to school was classified as within an acceptable criterion for active commuting (2.5 miles) or not. Physical Environment. The physical environment

Self-report questionnaire data were collected from 4720 adolescents in the Take PART study: Physical Activity Research for Teenagers, a cross-sectional study of the psychosocial and environmental determinants of physical activity and health among 15 to 17 year old adolescents in Ireland. Details on the rationale and methods are available elsewhere.27 All individuals signed an informed consent form, and parental consent was obtained for those under age 16. Ethical approval was granted by Dublin City University Ethics Board. Previous research has indicated that lengthy distances to school are associated with inactive modes, and that 2.5 miles is a realistic and achievable distance within which adolescents can actively commute.28 To remove distance as a confounding factor, this analysis is focused on individuals who live within 2.5 miles, providing a sample of n = 1143 males and n = 1016 females (47.1% female, mean age 16.04 ± 0.66). The sample was split between high (49.3%) and low (50.7%) socioeconomic status and urban (city: 19.1% and suburbs: 32.4%) and rural (town: 41% and village: 16%) residents. There was no gender difference in area of residence.

was measured using the Neighborhood Environment Walkability Scale (NEWS)31 and a convenient facilities scale.32 The NEWS measured resident’s perceptions of neighborhood characteristics thought to be related to the frequency of walking and cycling trips: pedestrian or traffic safety, safety from crime, aesthetics, proximity to stores and facilities (or land use mix diversity), perceived access to these destinations (or land use mix accessibility), places for walking and cycling (or function), neighborhood satisfaction, and street connectivity. Designed for use in a US setting, the NEWS has since been adapted (mostly with language/word changes to increase ecological validity) and used in Scotland,33 Australia,34 and Hong Kong35 as well as this study. Examples of wording changes include replacing ‘sidewalks’ with ‘pathways’ or ‘paths’ and removing inappropriate geographical features such as ‘canyons.’ Alpha reliability has been demonstrated for the original31 and adapted NEWS used in this study.27 The convenient facilities scale ascertained if facilities for sports (eg, running track or football field), exercise (eg, gym or aerobic dance studio), and general/ lifestyle physical activity (eg, bicycle lane or public park) were on a frequently traveled route (for example, to and from school) or within a 10-minute walk for their home. Words were changed to increase suitability to an Irish context, for example public recreation center was replaced with community center. The original (ICC = 0.80)32 and adapted version in this study (ICC = 0.60)27 demonstrated test-retest reliability.

Measures

Data Analysis

Methods Study Overview

Sociodemographics. Area of residence was classified

as i) large city (>500,000 inhabitants), ii) suburbs or outskirts of a city ( 50,000), iii) town ( 30).37,38 As simultaneous analysis of multiple interrelated variables could lead to difficulty distinguishing individual effects, it was decided to enter each summary score and each item in individual models. These logistic regression models examined the odds of active compared with inactive commuting to school. All models were run separately for males and females, and were adjusted for potentially influential sociodemographic factors (age and socioeconomic status). Variables found to be significant were included in a final multivariate model, adjusted for sociodemographics, density (size of settlement) and distance traveled to school (miles). For clarity, regression analyses results will be presented in 3 steps: (i) bivariate analyses with 9 summary score variables (8 NEWS subscales and the convenient facilities construct), ii) bivariate analyses with each item on each subscale (total of 80 items), and (iii) multivariate analyses with items significant in bivariate analyses. In the interests of space only significant results are tabulated.

Results Incidence of Active Commuting The majority of adolescents chose active modes of travel (61.3% walked and 8.7% cycled). Gender differences were due to differing proportions of cyclists and car users (χ2= 145.68, P < .001, r = .26); boys were more likely to cycle to school (15.4% vs. 1.2%) and girls were more likely to travel by car (27.0% vs. 18.3%). There was no gender difference in walking rates or numbers taking the bus/train.

Perceptions of the Physical Environment for Walking or Cycling Gender differences existed in perceptions of the neighborhood environment (Table 1). Males were more likely than females to perceive greater than 16 shops and amenities within a 10-minute walk of home, and more likely to report the presence of bike lanes and satisfaction with the ease of cycling in their local neighborhood.

Table 1 Descriptive Statistics for Subscale Summary Scores (Mean ± SD) and Specific Items (% Agreeing With Statement) by Gender Environmental variable Pedestrian/Traffic Safety (Range 8–32) Pedestrian crossings-present Safety from crime (Range 7–28) Unsafe to walk day Unsafe to walk night Aesthetics (Range 6–24) Interesting features Attractive natural sights Land-use mix diversity (Range 20–100) 16–20 amenities within 10 min of home Land-use mix access (Range 7–28) Can do most shopping at local shops Shops within easy walking distance Parking difficult in local shopping areas Easy to walk to public transit Streets are hilly Convenience of PA facility (Range 17–51) Bike Lane Sea/beach Places for walking and cycling (Range 6–24) Neighborhood satisfaction (Range 4–20) Ease of cycling Connectivity (Range 5–20) Walkways connect cul-de-sacs Crossroads close together

Males (n = 1143)

Females (n = 1016)

21.51 ± 3.91 53.8 21.32 ± 3.67 12.9 32.8 14.38 ± 3.73 34.7 34.8 68.73 ± 18.00 16.2 21.17 ± 3.92 69.8 78.4 54.3 64.9 74.7 36.30 ± 6.55 52.4 38.2 17.36 ± 3.59 15.92 ± 3.55 70.6 12.06 ± 2.78 42.1 46.5

21.52 ± 4.03 61.2 21.26 ± 3.48 9.1 39.8 14.79 ± 3.78 40.9 40.3 65.53 ± 17.70 12.3 22.26 ± 3.93 75.8 87.3 63.5 74.9 82.4 36.55 ± 6.18 47.6 29.8 17.44 ± 3.73 15.94 ± 3.36 60.5 12.34 ± 2.71 50.1 53.5

P ** * ** * ** ** ** ** *** * *** ** *** ** * ***

*** * *

Note. Figures in bold are significantly higher as follows: * P < .05; ** P < .01; *** P < .001; analyzed using t tests for the summary scores and Pearson χ2 tests for the individual items. Only items where gender differences existed are included.


A higher proportion of girls, than boys, agreed that pedestrian crossings were present, that it was unsafe to walk at night, and that more interesting features and attractive natural sights existed in their local environment. Girls were more likely to perceive good accessibility to local shops, public transport, and presence of hilly streets. While no gender difference existed in perceptions of places for walking or cycling, girls were more likely to report good connectivity in terms of walkways connecting cul-de-sacs and having crossroads close together.

Bivariate Analyses With Subscale Scores In general, positive perceptions of places for walking/ cycling, land use mix diversity and land use mix access increased the odds of active commuting to school (Table 2). Perceptions of aesthetics and convenient physical activity facilities were related to AC among males only. Four subscales were unrelated to active travel behavior: specifically pedestrian/traffic safety, safety from crime, neighborhood satisfaction, and connectivity.

Bivariate Analyses With Items—Males Perceptions of specific environmental features were related to increased odds of active commuting to school among boys (Table 2). Visibility was important; males who agreed that streets are well lit, that walkers are seen by others, and that they see and speak to others while walking in their neighborhood were more likely to actively commute to school. Positive perceptions of land use mix diversity and access were also related; perceptions of more than 5 amenities within a 5-to-10-minute walk of home increased the odds of active commuting. Boys who perceived good access to shops, public transport, had ‘places to go’ and convenient bike lanes, public parks, and the sea/beach had increased the odds of walking or cycling to school. Perceived presence and maintenance of paths that were separated from the road (either by parked cars or grass verges) and perceptions of walkways connecting cul-de-sacs increased the likelihood of male active travel. Fewer features were associated with decreased odds of active commuting. Boys who perceived interesting features or attractive natural sights in their neighborhood were less likely to walk or cycle to school. Perceptions of litter free streets were also linked with reduced odds of active commuting to school. Pedestrian safety was not associated with mode of travel to school for boys. A number of individual items from other subscales were also unrelated to active travel including crime rate, the presence of trees and attractive buildings, the number of convenient PA facilities, accessible cycle lanes and ease of walking and cycling.

Bivariate Analyses With Items—Females In contrast to findings among males, many features of traffic safety were associated with girl’s mode choice (Table 2). Girls who agreed that pedestrian crossings were present, and who perceived that pedestrian crossings increased safety in their neighborhood, had increased odds of actively commuting to school. In addition, girls who agreed that traffic speed was slow were more likely to actively commute. Perceptions of safety from crime were less important for females than males, however females who agreed that streets are well lit were more likely to actively commute to school. Similar to results among males, positive perceptions of land use mix diversity, land use mix access, convenient facilities, and function were related to active commuting among females. In addition, 3 items that were unimportant for males supported active travel among girls. Perceptions of easily accessible cycle paths and trees offering shelter to walkers were associated with increased odds of active commuting. In addition, girls who were satisfied with the ease of walking in their local neighborhood were more likely to actively commute to school. None of the individual items measuring street connectivity were related to girl’s active travel behavior. Also unrelated were traffic density, crime rates, walkers seen by others, attractive natural sights and buildings, hilly streets, and the number of convenient PA facilities within 5 to 10 minutes of home.

Multivariate Analyses Items that remained significant in the multivariate models for males are presented in Table 3. When the model was adjusted for density, perceptions of footpaths, interesting features, and shops within walking distance were no longer significant predictors of active commuting to school. In addition, when the model was adjusted for the effect of distance on commuting behavior, perceptions of attractive natural sights and having paths separate from the road were no longer significant. The addition of distance also removed the effect of population density on active commuting to school, indicating an interaction between these variables. Items that remained significant in the multivariate model for females are presented in Table 4. A large number of variables were no longer significant, for example perceptions of traffic speed and convenient physical activity facilities. After controlling for density, the effect of perceptions of pedestrian crossings and the presence of paths was no longer significant. Similar to results among males, the effect of population density on active commuting to school was no longer significant when distance was added to the model.

Table 2 Associations Between Perceptions of Environment and Active Commuting to School for Males and Females aOR (95% CI) Environmental variable Pedestrian/Traffic Safety Traffic speed slow my street Traffic speed slow nearby streets Drivers exceed speed limits Pedestrian crossings—present Exhaust fumes Pedestrian crossings—feel safe Safety from crime Streets well lit at night Walkers seen by others See and speak to others Unsafe to walk at night Aesthetics Trees give shelter Interesting features Free from litter Attractive natural sights Land-use mix diversity # within 5 min of home: 0–5 6–10 11–15 16–20 # within 10 min of home: 0–5 6–10 11–15 16–20 Land-use mix accessibility Shops within easy walking distance Many places to go Easy to walk to public transit Streets are hilly Convenience of PA facility Bike Lane Public park Sea/beach Places for walking/cycling Paths present Paths well maintained Cycle lanes/paths are easy to get to Cars separate paths from road Grass separates paths from roads Neighborhood satisfaction Ease of walking Connectivity Walkways connect cul-de-sacs

1.01 1.15 1.16 1.04 1.25 1.12 1.15 1.02 1.41 1.40 1.64 1.29 .93 1.01 .65 .64 .58 1.05

Male (n = 1143) (.98-1.05) (.87-1.52) (.89-1.52) (.79-1.36) (.96-1.64) (.85-1.47) (.88-1.51) (.98-1.05) (1.05-1.88) * (1.06-1.84) * (1.23-2.20) *** (.96-1.73) (.90-.97) *** (.77-1.32) (.49-.83) ** (.49-.83) *** (.44-.76) *** (1.03-1.06) ***

1.06 1.48 1.38 0.63 1.77 1.32 1.62 1.00 1.47 1.22 1.35 1.43 .97 1.31 .97 .62 .85 1.04

Female (n = 1016) (1.03-1.10) (1.13-1.94) *** (1.06-1.79) * (.48-.83) *** (1.36-2.31) *** (1.01-1.71) * (1.24-2.10) *** (.96-1.04) (1.11-1.94) ** (.93-1.60) (.99-1.84) (1.09-1.88) ** (.94-1.01) (1.01-1.71) * (.74-1.26) (.48-.81) *** (.65-1.10) (1.02-1.05) ***

1.0 2.48 3.42 3.24

(1.28-4.81) ** 1.26-9.24) * 1.45-7.23) **

1.0 3.19 2.86 1.27

(1.68-6.05) *** (1.08-7.53) * (.58-2.78)

1.0 4.74 5.40 8.01 1.14 3.74 2.22 2.33 1.66 1.03 1.38 1.56 1.59 1.10 3.60 1.44 1.26 1.48 1.75 .98 1.38 1.06 1.83

(2.59-8.67) *** (2.88-10.11) *** (4.41-14.53) *** (1.08-1.20) *** (2.28-6.12) *** (1.43-3.45) *** (1.52-3.57) *** (1.05-2.63) * (1.01-1.05) ** (1.06-1.80) * (1.14-2.13)** (1.16-2.16) ** (1.06-1.15) *** (2.38-5.45) *** (1.06-1.97) * (.94-1.68) (1.10-1.97) ** (1.31-2.34) *** (.94-1.02) (.84-2.29) (.98-1.14) (1.19-2.82) **

1.0 3.54 5.0 3.01 1.13 10.28 2.68 2.08 1.34 1.02 1.75 1.37 1.98 1.12 4.18 1.56 1.62 1.81 1.40 .98 2.19 1.06 1.14

(2.04-6.14) *** (2.81-8.89) *** (1.81-4.99) *** (1.08-1.19) *** (4.84-21.87) *** (1.75-4.09) *** (1.34-3.21) *** (.83-2.18) (.99-1.04) (1.32-2.24) *** (.98-1.90) * (1.41-2.78) *** (1.07-1.17) *** (2.61-6.68) *** (1.14-2.12) ** (1.21-2.19) ** (1.34-2.43) *** (1.04-1.88) * (.94-1.03) (1.24-3.86) ** (.98-1.13) (.78-1.65)

Note. aOR = odds ratios adjusted for age and socioeconomic status. Only significant items reported. * P < .05. ** P < .01. *** P < .001.

261

Table 3 Multivariate Model for Associations Between Perceptions of the Physical Environment and Active Commuting to School Among Males aOR (95% CI) Environmental variable Aesthetics Interesting features Attractive natural sights Land-use mix diversity # within 10 min of home 0–5 6–10 11–15 16–20 Land-use mix accessibility Shops easy walking distance Many places to go Convenience of PA facility Public park Places for walking/cycling Paths present Cars separate paths from road Density City Suburbs Town Village Distance

Sociodemographics

Density

Distance

0.65 (0.45-0.96) * 0.42 (0.29-0.62) ***

0.72 (0.49-1.08) 0.48 (0.32-0.71) ***

0.77 (0.49-1.22) 0.78 (0.42-1.04)

1.0 3.89 (2.37-6.37) *** 3.29 (1.88-5.76) *** 5.14 (2.92-9.07) ***

1.0 3.88 (2.33-6.45) *** 2.87 (1.60-5.12) *** 4.10 (2.28-7.38)***

1.0 2.89 (1.61-5.17) ** 1.69 (0.86-3.30) 2.11 (1.06-4.22) *

1.55 (1.03-2.35) * 1.49 (1.01-2.21) *

1.44 (0.94-2.20) 1.43 (0.96-2.14)

1.10 (0.75-1.89) 1.25 (0.76-2.04)

2.18 (1.49-3.20) ***

1.90 (1.28-2.83) ***

1.63 (1.03-2.59) *

1.66 (1.01-2.73) * 1.49 (1.03-2.16) *

1.43 (0.85-2.39) 1.51 (0.13-2.21) *

1.21 (0.66-2.21) 1.50 (0.97-2.34)

—– —– —– —– —–

18.16 (2.97-10.85)*** 6.38 (2.75-14.77) *** 3.01 (1.99-4.56) *** 1.0 —–

8.56 (0.78-9.56) 2.03 (0.80-5.15) 1.14 (.86-2.30) 0.53 (0.46-0.61) ***

Note. The reference category is always in disagreement with the statement. * P < .05. ** P < .01. *** P < .001.

Table 5 summarizes the specific features that support and hinder active commuting behavior among adolescents in this study compared with previous research.

Discussion Understanding the specific environmental characteristics that influence active commuting to school is essential for effective intervention. This study is among the first to examine the relationships between perceptions of the physical environment and active commuting to school among adolescents. In support of previous research,16 specific environmental features were positively associated (possible supports), while others were negatively associated (possible barriers) with active commuting. Although these findings contribute to understanding of the environment-physical activity relationship, more research is required before strong recommendations can be made for intervention. The results illustrate a number of issues and challenges in this field of inquiry that must be addressed before research can inform effective intervention and evidence-based policy. This discussion highlights 3 issues: gender differences in associations,

262

measurement of perceived environment, and collinearity in data analysis. This study suggests that associations between the perceived environment and active travel behavior can vary by population subgroup. Only 2 previous studies have addressed gender as a potential mediator of environmental influences on walking and cycling among young people. Evenson and colleagues (2006) sampled adolescent girls only while Carver and colleagues (2005) compared adolescent boys and girls. Similar to these studies, traffic safety,17 visibility, the presence of cycle tracks, and the ease of walking/cycling to transit39 were related to girl’s active travel in this study. Contrary to previous research,17 traffic density was not associated with walking or cycling to school for girls. In addition, the number39 and quality17 of sports facilities were positively associated with girl’s active travel in previous studies, but perceptions of many convenient physical activity facilities was unrelated to active travel in this sample. A number of items that were unrelated to active commuting in previous research39 were associated with active commuting in this study, including well lit streets, trees offering shelter, litter-free streets, the presence of paths, and land use mix diversity

Environment and Active Commuting 263

Table 4 Multivariate Model for Associations Between Perceptions of the Physical Environment and Active Commuting to School Among Females aOR(95% CI) Environmental variable Pedestrian/traffic safety Drivers exceed speed limits Pedestrian crossings-present Safety from crime Unsafe to walk at night Aesthetics Free from litter Land-use mix diversity # within 10 min of home 0–5 6–10 11–15 16–20 Land-use mix accessibility Shops within easy walking distance Easy to walk to public transit Places for walking/cycling Paths present Cars separate paths from road Density City Suburbs Town Village Distance

Sociodemographics

Density

Distance

0.65 (0.46-0.93) * 1.46 (1.00-2.12) *

0.65 (0.45-0.94) * 1.34 (0.91-1.97)

0.62 (0.40-0.94) * 1.23 (0.79-1.92)

1.91 (1.31-2.79) **

1.99 (1.34-2.95) **

1.98 (1.25-3.14) **

0.54 (0.38-0.78)**

0.57 (0.39-0.82) **

0.60 (0.39-0.93)*

1.0 1.87 (1.14-3.06) * 1.93 (1.17-3.19) * 2.78 (1.51-5.12) **

1.0 1.78 (1.07-2.97) * 1.85 (1.10-3.10) * 2.19 (1.16-4.13) *

1.0 1.77 (0.98-3.18) 1.76 (0.96-3.22) 1.07 (0.52-2.19)

5.82 (3.06-11.07) *** 1.99 (1.32-2.98) ***

5.30 (2.71-10.36) *** 1.63 (1.06-2.50) *

4.27 (2.03-8.95) *** 1.58 (0.97-2.55)

1.87 (1.08-3.24) * 1.52 (1.06-2.19) *

1.58 (0.89-2.81) 1.56 (1.07-2.27) *

1.30 (0.67-2.49) 1.63 (1.06-2.50) *

—– —– —– —– —–

18.92 (5.34-67.01) *** 4.68 (2.32-9.44) *** 3.88 (2.52-5.97) *** 1.0 —–

3.36 (0.83-13.35) 1.49 (0.66-3.33) 1.55 (0.91-2.63) 1.0 0.51 (0.44-0.59) ***

Note. The reference category is always disagreement with the statement. * P < .05. ** P < .01. *** P < .001.

and access. Despite methodological similarities between these studies, both sampled younger adolescents, aged 12 to 15 years, which may account for different findings. In some cases, gender differences in factors associated with active commuting can be intuitively explained. For example, girls traditionally have less independent mobility than boys40,41 which might result in increased fear and learned helplessness surrounding navigation of the physical environment. This increased vulnerability might explain why girls are more affected by traffic safety issues. In other cases, gender differences in the impact of environmental perceptions on active commuting behavior might be explained by underlying differences in perceptions. For example, females are more likely to agree that walkways connect cul-de-sacs in their neighborhood, but this environmental characteristic does not influence their active travel behavior. Boys, who are less likely to report the presence of walkways, are more likely to walk or cycle to school when they do exist. This suggests a

different effect of perceived absence of positive features and perceived presence of positive features. While Carver and colleagues (2005) identified gender differences in perceptions, and the effect of perceptions on walking and cycling behaviors, the mechanisms through which this effect might occur were not examined.17 This is symptomatic of a slow move toward trying to merge understanding of environmental correlates with individual cognitions and behavioral intention or change.42 The second issue identified in this study relates to the use of survey measures and their subsequent scoring and presentation. The results suggest that the level of measurement may influence interpretation of effects. Using the NEWS questionnaire, 5 subscales were important predictors of active commuting to school among males: aesthetics, land use mix diversity, land use mix access, convenient physical activity facilities, and function; and 3 were significant for females: land use mix diversity and access, and convenient physical activity facilities. However, analysis by item revealed


Table 5 Specific Features That Support and Inhibit Active Commuting Behavior Among Adolescents Supports active commuting

Inhibits active commuting

Slow traffic speedf Pedestrian crossings presentf Pedestrian crossings safef Well lit streets√ Walkers visiblem,√ Trees offer shelterf Number of amenities (5–10 min walk) Shops (easy walking distance) Many places to go (easy walking distance) Public transport (easy walking distance) Convenient bike lane Convenient public park√ Convenient sea/beach Footpaths present Footpaths well maintained Cycle lanes easy to get tof Footpath separate from road by parked cars or grass verges Easy to walkf,√ Counter-intuitive findings Exhaust fumes presentf,X Unsafe to walk at nightf Hilly streets m

Drivers exceeding speed limitsf,√ Counter-intuitive findings: Interesting features to look atm,X Litter free streetsX

Note. Gender specific findings marked. Bold emphasis indicates variables that remained significant in multivariate analysis. f Females. m Males. √ Findings that support previous research. X Findings that contradict previous research.

that specific items from the other scales: pedestrian/ traffic safety, safety from crime, neighborhood satisfaction, and connectivity, were also important predictors of active commuting. It appears that reliance on subscale summary scores risks the loss of important information. As a result, where possible, research should focus on perceptions of specific characteristics of the environment (such as ‘the presence of pedestrian crossings’) rather than overall perceptions (such as ‘perceptions of pedestrian safety’). A third challenge for researchers is to address issues related to multicollinearity of environmental variables and how this impacts on data analysis. Environmental characteristics are unlikely to be completely independent and conducting numerous individual analyses presents only an exploratory picture of effects. Multivariate analyses can provide information on the relative contribution of each feature however where issues around multicollinearity exist, the simultaneous analysis of multiple constructs that are interrelated may yield spurious or confounded results whereby it is impossible to distinguish the individual effects. In the multivariate model presented, land

use mix diversity and the perceived presence of a public park remained associated with active commuting for males and perceptions of excess traffic speed, access to shops within walking distance and paths separate from the road remained associated with active commuting for females. However, given the suggested collinearity between the variables included, it cannot be determined if the variables that remain associated with active commuting have individual or interactive effects. Future research should consider factor analyses43 or multilevel modeling with interactions between environmental features, such as structural equation modeling.42

Strengths and Limitations This piece of work is similar to other early investigations into environmental influences on physical activity behavior—a multifaceted phenomenon that requires continued study.44 This study has measured only selfreported perceptions of the environment and cannot make recommendations for intervening with actual environments. Moreover, the methods and results suggest ways

Environment and Active Commuting 265

to improve current research before making recommendations for intervention and policy. The strengths and novelty of this study include a substantial sample size and variable neighborhood environments provided by sampling urban and rural areas of residence. A wide variety of environmental variables were measured and many additional variables were included that were not previously measured in this age group. In addition, using a distance criterion to ensure inclusion of only adolescents who could realistically walk or cycle to school strengthened this analysis. As well as controlling for the potentially confounding effect of distance, all analyses controlled for socioeconomic status. As the environment is hypothesized to influence different behaviors in different ways, the behavior-specific nature of this analysis offers advantages over previous more general studies. This study is limited by its reliance on cross-sectional data, resulting in an inability to infer causal relations; it is unknown whether perceptions of environmental features cause active commuting or whether active commuters are more aware of these features and therefore have altered perceptions. It is fundamentally important to accept that the examination of items as individual predictors, although a limitation, is necessary to develop concepts around the relevance of specific features. This approach has been applied by other researchers in the field39 and the results provide a clear pathway for future research. Confirming the demonstrated associations in similar studies, and in other populations and countries, a clearer list of variables to target in interventions can be developed.

Recommendations for Future Research Based on the experiences of this study, recommendations for future research include: (i) more focus by researchers on understanding the measurement and analysis of environmental variables, (ii) the use of NEWS items in analyses rather than summary scores or the use of factor analyses to examine the stability of the NEWS subscales as constructs, (iii) multilevel modeling allowing for interactions between environmental features, (iv) further examination of gender as a mediator or moderator of environmental impacts, (v) consideration of the theoretical basis linking perceived environments to behavioral intentions to behavioral change, (vi) interventions addressing links between perceptions and behavior, and (vii) interventions addressing actual environmental change.

Conclusions The relationship between the physical environment and physical activity among youth has been the emphasis of recent research16 however there is less behavior-specific data on the environmental correlates of walking and cycling to school.19 In support of previous research,16 specific environmental features were positively associated (possible supports), while others were negatively associated (possible barriers) with active commuting.

Although these findings contribute to understanding of the environment-physical activity relationship, more research is required before strong recommendations can be made for intervention. The results illustrate a number of issues and challenges relating to gender differences in associations; measurement of perceived environment and collinearity in data analysis must be addressed before research can inform effective intervention and evidencebased policy.

References 1. Strong WB, Malina RM, Blimkie CJ, et al. Evidence based physical activity for school-age youth. J Pediatr. 2005;146(6):732–737. 2. World Health Organisation. Global strategy on diet, physical activity and health; 2004. 3. Sallis JF. Age-related decline in physical activity: a synthesis of human and animal studies. Med Sci Sports Exerc. 2000;32(9):1598–1600. 4. Telema R, Yang X. Decline of physical activity from youth to young adulthood in Finland. Med Sci Sports Exerc. 2000;32(9):1617–1622. 5. Caspersen CJ, Pereira MA. Curran, K. M. Changes in physical activity patterns in the United States, by sex and cross-sectional age. Med Sci Sports Exerc. 2000;32(9):1601–1609. 6. Cooper AR, Andersen LB, Wedderkopp N, Page AS, Froberg K. Physical activity levels of children who walk, cycle or are driven to school. Am J Prev Med. 2005;29(3):179– 184. 7. Cooper AR, Page AS, Foster LJ, Qahwaji D. Commuting to school: are children who walk more physically active? Am J Prev Med. 2003;25(4):273–276. 8. Sirard JR, Riner WF, Jr, McIver KL, Pate RR. Physical activity and active commuting to elementary school. Med Sci Sports Exerc. 2005;37(12):2062–2069. 9. Tudor-Locke C, Ainsworth BE, Adair LS, Popkin BM. Objective physical activity of Filipino youth stratified for commuting mode to school. Med Sci Sports Exerc. 2003;35(3):465–471. 10. Nelson NM. (2007) The health impact and environmental determinants of active travel to school. PhD Thesis. School of Health and Human Performance, Dublin City University. 11. Cooper AR, Wedderkopp N, Wang N, Bo Andersen L, Froberg K, Page AS. Active travel to school and cardiovascular fitness in Danish children and adolescents. Med Sci Sports Exerc. 2006;38(10):1724–1731. 12. Brodersen NH, Steptoe A, Williamson S, Wardle J. Sociodemographic, developmental, environmental and psychological correlates of physical activity and sedentary behaviour at age 11 to 12. Ann Behav Med. 2005;29(1):2–11. 13. Booth SL, Sallis JF, Ritenbaugh C, et al. Environmental and societal factors affect food choice and physical activity: Rationale, influences and leverage points. Nutr Rev. 2001;59(3):S21–S39. 14. Giles-Corti B, Timperio A, Bull F, Pikora T. Understanding physical activity environmental correlates: increased specificity for ecological models. Exerc Sport Sci Rev. 2005;33(4):175–181. 15. Sallis JF, Owen N, Glanz K, Lewis FM, Rimer BK. Ecological Models. In: Health Behaviour and Health Education. San Francisco: Jossey-Bass Publishers; 1997:403–424.


16. Krahnstoever Davison K, Lawson CT. Do attributes in the physical environment influence children’s physical activity? A review of the literature. Int J Behav Nutr Phys Act. 2006;3(19). 17. Carver A, Salmon J, Campbell K, Baur L, Garnett S, Crawford D. How do perceptions of local neighbourhood relate to adolescents’ walking and cycling? Am J Health Promot. 2005;20(2):139–147. 18. Ewing R. Can the physical environment determine physical activity levels? Exerc Sport Sci Rev. 2005;33(2):69–75. 19. Panter J, Jones AP, van Sluijs EMF. Environmental determinants of active travel in youth: a review and framework for future research. Int J Behav Nutr Phys Act. 2008;5:34. 20. Frank L, Kerr J, Chapman J, Sallis J. Urban form relationships with walk trip frequency and distance among youth. Am J Health Promot. 2007;21:305–311. 21. Hohepa M, Scragg R, Schofield G, Kolt G, Schaat D. Social support for youth physical activity: Importance of siblings, parents, friends and school support across a segmented school day. Int J Behav Nutr Phys Act. 2007;4:54. 42. Ewing R, Schroeer W, Greene W: School location and student travel analysis of factors affecting mode choice. Transportation Research Board: Journal of the Transportation Research Board 2004:55-63. No 1895 TRB 22. Timperio A, Crawford D, Telford A, Salmon J. Perceptions about the local neighbourhood and walking and cycling among children. Prev Med. 2004;38:39–47. 23. Timperio A, Ball K, Salmon J, et al. Personal, family, social, and environmental correlates of active commuting to school. Am J Prev Med. 2006;30:45–51. 24. Centers for Disease Control and Prevention. Barriers to children walking to and from school – United States. JAMA. 2002;51(32):701–704. 25. Central Statistics Office, Government of Ireland. Census of Population Volume 9: Travel to work, school and college. 2002; Available from: www.cso.ie 26. Department of Transport UK. Transport Statistics Bulletin. National Travel Survey 2005: National Statistics; 2006. 27. Woods CB, Nelson NM, Foley E, O’Gorman D, Moyna NM. The Take PART study (Physical Activity Research for Teenagers): rationale and methods. J Phys Act Health. 2009;6:170–177. 28. Nelson NM, Foley E, O’Gorman D, Moyna NM, Woods CB. Active commuting to school: how far is too far? Int J Behav Nutr Phys Act. 2008;5(1). 29. Currie CE, Samdal O, Boyce W, Smith R. Health Behaviour in School-Aged Children: A WHO Cross-National Study (HBSC), Research Protocol for the 2001/2002 Survey. University of Edinburgh: Child and Adolescent Health Research Unit. CAHRU; 2001. 30. Duncan GJ, Brooks-Gunn J. Family Poverty, Welfare Reform, and Child Development. Child Dev. 2000;71(1):188–196.

31. Saelens BE, Sallis JF, Frank LD. Environmental correlates of walking and cycling: findings from the transportation, urban design, and planning literatures. Ann Behav Med. 2003;25(2):80–91. 32. Sallis JF, Johnson MF, Calfas KJ, Caparosa S, Nichols JF. Assessing perceived physical environmental variables that may influence physical activity. Res Q Exerc Sport. 1999;68(4):345–347. 33. Fitzsimons CF, Baker G, Wright A, et al. The ‘Walking for Wellbeing in the West’ randomised controlled trial of a pedometer-based walking programme in combination with physical activity consultation with 12 month follow-up: rationale and study design. BMC Public Health. 2008;8:259. 34. Cerin E, Leslie E, Owen O, Baumann A. An Australian version of the neighbourhood environment walkability scale: validity evidence. Meas Phys Educ Exerc Sci. 2008;12(1):31–51. 35. Cerin E, Macfarlane DJ, Ko H-H, Chan K-CA. Measuring perceived neighbourhood walkability in Hong Kong. Cities. 2007;24(3):209–217. 36. Pett MA. Nonparametric Statistics for Health Care Research: Statistics for Small Samples and Unusual Distributions. Sage Publications Inc.; 1997. 37. Tabachnick BG, Fidell LS. Using Multivariate Statistics. Pearson International; 2007. 38. Field A. Discovering Statistics using SPSS. SAGE Publications; 2005. 39. Evenson KR, Birnbaum AS, Bedimo-Rung AL, et al. Girls’ perception of physical environmental factors and transportation: reliability and association with physical activity and active transport to school. Int J Behav Nutr Phys Act. 2006;3(28). 40. Day K, Boarnet M, Alfonzo M, Anderson C. Johnny walks to school - does Jane? Sex differences in children’s active travel to school. Children. Youth and Environments. 2006;16:75–89. 41. Prezza M, Pilloni S, Morabito C, Sersante C, Alparone Romana F, Vittoria Giuliani M. The influence of psychosocial and environmental factors on children’s independent mobility and relationship to peer frequentation. J Comm and App Soc Ppsych. 2001;11:435–450. 42. Nelson NM, Wright A, Lowry RG. Mutrie, N. Where is the theoretical basis for understanding and measuring the environment for physical activity? Environmental Health Insights. 2008;2:111–116. 43. Cerin E, Saelens BE, Sallis JF, Frank LD. Neighbourhood environment walkability scale: validity and development of a short form. Med Sci Sports Exerc. 2006;38(9):1682–1691. 44. Norman GJ, Nutter SK, Ryan S, Sallis JF, Calfas KJ, Patrick K. Community design and access to recreational facilities as correlates of adolescent physical activity and body-mass index. J Phys Act Health. 2006;3(S1):S118–S128.