PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Adolesc Health. Author manuscript; available in PMC Nov 1, 2011.
Published in final edited form as:
PMCID: PMC2963857
NIHMSID: NIHMS192943
The Association between Community Physical Activity Settings and Youth Physical Activity, Obesity and BMI
Sandy J. Slater, PhD,1 Reid Ewing, PhD,2 Lisa M. Powell, PhD,1 Frank J. Chaloupka, PhD,1 Lloyd D. Johnston, PhD,3 and Patrick M. O’Malley, PhD3
1University of Illinois at Chicago, Institute for Health Research and Policy, Chicago, Illinois
2University of Utah, Department of City and Regional Planning
3University of Michigan, Survey Research Center, Institute for Social Research, Ann Arbor, Michigan
Keywords: adolescents, environment, exercise
The U.S. has been experiencing a growing trend in overweight and obesity among youth and recent evidence shows that approximately 32 percent of youth are overweight or obese [13]. Physical inactivity impacts weight and in an effort to curb the growing obesity epidemic there is an increasing body of behavioral physical activity (PA) research that has examined associations between local area environmental factors and PA among youth [4, 56]. Greater availability of outdoor play/sports areas and parks [711], and access to commercial PA-related facilities [10, 1215] have been associated with higher levels of youth PA. However, perceived environmental barriers, such as lack of access to these types of settings, have been associated with lower income neighborhoods [1617].
Neighborhood design can also influence PA levels in youth. Neighborhoods with more cul-desacs or low connectivity street networks are associated with higher levels of home-based PA in youth than are high walkability (grid street network) neighborhoods, but these same neighborhoods are associated with lower levels of active travel [18]. More sprawling neighborhoods have been connected with higher prevalence of overweight among youth [4, 6]. Issues of safety, such as danger from traffic, stranger danger, and perceptions of safety have all been found to have a negative impact on PA levels [5, 10, 19].
The purpose of this study was to determine whether there is an association between the level of PA friendliness of the built environment (i.e., more compact communities with less traffic danger and physical disorder, and the presence of both commercial and outdoor PA settings) and youth PA and weight. This study used a national sample of 8th and 10th grade students (13–16 years old) to simultaneously examine the association between measures of neighborhood PA settings, safety, and community compactness/walkability and youth PA, obesity, and BMI. Although other studies have examined these environmental factors individually, they could have been subject to omitted variables biases. This study’s key strength is that it is the first to simultaneously examine the association of these multiple factors on PA and weight at the national level using environmental data collected directly from the youths’ neighborhoods.
This study combined repeated cross-sections from 2001–2003 of individual-level national data for 8th and 10th grade students participating in the Monitoring the Future (MTF) survey with community-level environmental measures developed through the ImpacTeen Project (www.impacteen.org).
Monitoring the Future Survey Data
Since 1991, the MTF study, a survey of youth substance use and abuse, conducted by the University of Michigan’s Institute for Social Research and funded by the National Institute on Drug Abuse, has collected annual national samples of about 33,000 8th and 10th grade students in the coterminous United States. Located in approximately 420 schools, these students/schools were selected based on a three-stage sampling procedure, which is described in detail elsewhere [20]. Questionnaires were administered by an ISR representative in classrooms during normal class periods whenever possible. Students were informed of the importance of accurate responses and assured that their confidentiality would be protected.
MTF Physical Activity and Weight Outcome Measures
PA measures were based on self-reported responses to three questions: 1) “How often do you exercise vigorously (jogging, swimming, calisthenics, or any other active sports)?” (vigorous exercise), 2) “To what extent did you participate in school athletic teams this school year?” (sports participation), and 3) “How often do you do actively participate in sports, athletics or exercising?” (PA participation). Using self-reported height and weight, we calculated BMI (=weight(kg)/height(m)2). Individuals’ body weight status was classified based on BMI for children and teens using the 2000 CDC Growth Chart [21], obesity was classified as BMI>=age-sex-specific 95th percentile.
ImpacTeen Community Data
In any given year, half the participating MTF schools are in their first year of participation and half are in their second year. Only the half-sample of schools cycling out of the MTF survey were involved in the ImpacTeen community data collection activities. They constituted a national sample replicate. For the three years of data in this study there are approximately 12,000–24,000 students per grade per year for each outcome variable and 396 catchment areas.
Community observations were conducted each year beginning in mid-February and ending in early June. Environmental information was collected in conjunction with tobacco and alcohol marketing information in retail establishments. For each school, a catchment area was defined, reflecting the area from which the school drew the majority of its students, i.e. the school enrollment zone. Our sample size for this study consists of 134, 135 and 127catchment areas in 2001, 2002 and 2003 respectively.
Using the maps of the catchment areas as guides, trained field teams composed of a lead and an assistant observer were instructed to drive/walk around and record information on advertising, recreational space, social interactions, public events/signage, safety, and the general upkeep of the catchment area in order to characterize what is was like to live in the area. The two-person teams served as a reliability check for each site observation. If there was disagreement on any measure, field staff would return to verify the measure prior to leaving the community.
Community Environmental Measures
Because our outcome measures ask about more general PA and do not capture walking behavior, we have included measures of destinations where youth are likely to be physically active (commercial facilities and outdoor settings). We also included the local compactness index and measures of safety to explore how some forms of access to these settings might encourage or discourage PA behavior.
Using existing ImpacTeen data we constructed the following measures. A physical disorder scale was constructed and includes measures on the presence/absence of: 1) homeless persons loitering on the streets; 2) bars on windows of buildings; 3) unkempt/dilapidated buildings; 4) security barriers around residential and retail property; 5) teens hanging out (showing negative behaviors such as smoking or drinking); and, 6) vandalism and/or graffiti (Cronbach’s Alpha (CA)=0.75). An outdoor PA-related settings scale was constructed and includes measures on the presence/absence of: 1) sports areas (e.g. baseball diamonds, basketball and tennis courts, soccer fields, etc.); 2) parks/green spaces, playgrounds, golf courses; 3) public pools/beaches; and 4) presence of bike paths/lanes (CA=0.60).
Data on the availability of commercial PA-related outlets (PA outlets) were obtained from a business list developed by Dun and Bradstreet [22] to construct a measure of PA-related facilities based on 100 different 8-digit standard industry classification (SIC) codes, such as physical fitness facilities, sports and athletic instruction (i.e. gymnastics), YMCA, etc. The total number of outlets were summed within zip code and divided by zip code population expressed in 10,000 to develop a measure of PA outlet availability per 10,000 capita. The PA-related outlet density measure was matched to the individual-level data at the school zip code level.
A measure of perceived safety was drawn from the individual-level MTF self-report surveys where students were asked “How often do you feel unsafe going to or from school?”. Students’ responses were based on a 5-point scale that included: never, rarely, some days, most days, and every day. The question is form-specific. Therefore we constructed an aggregate school-level measure of student responses representing the proportion of students from each school who answered some days, most days and every day.
Using methods developed by Forsyth et al. [23], we constructed a measure of traffic danger using US Census TIGER file road classifications which assigns streets higher or lower functional classes. High road classifications include primary highways and roads, secondary and connecting roads, and roads that provide access to businesses, and facilities. The measure represents the ratio of higher road classes to local and neighborhood roads within the catchment area, i.e. the ratio of roads with higher speed limits and traffic volumes to all other roads.
Using existing methods [24] a local urban compactness index was developed for each catchment area including dimensions of residential density and street connectivity. A factor was extracted through principal components analysis to represent the level of compactness within a community (CA=0.82). We used 2000 U.S. Census data [25], matched at the census block level, to construct a measure of residential density per square mile of the catchment area and two measures for street connectivity (intersection density per square mile and ratio of 4-way intersections to all other intersections in the catchment area) using ArcGIS 9 software. The index was transformed to a scale with a mean of 100 and a standard deviation of 25. The larger the value of the index, the more compact the catchment area.
Control Measures
We controlled for basic demographic measures including: gender; grade; race/ethnicity; highest level of schooling completed by father and mother; inflation-adjusted total weekly student income (earned and unearned, such as allowance); whether students worked; and mother’s work status. We controlled for region, year of data collection, whether the school was private vs. public, and neighborhood wealth effects by including a site-level measure of median household income from the Census 2000 [25].
Because students were clustered within sites, we ran a series of 2-level random intercept regression models using SAS glimmix and proc mixed to estimate the association between the built environment and outcome variables. The models adjusted the standard errors to account for within-community clustering, and appropriately modeled variance due to individual differences versus between-community differences (the intra-class correlation coefficient, or ICC) [26]. Sampling weights were used to adjust for differential selection probabilities for the schools. We tested relationships between individual student-level outcomes, individual-level predictors unique to each student, and community-level predictors shared by multiple students within communities.
The ICC was calculated as the area level variance / (area level variance + π2 /3) for logistic regression models and as the area level variance /(area level variance + residual) for BMI. Proportionate reduction in error (PRE) was calculated using the following formula: area-level variance of the conditional means model - area-level variance of model 2, 3, or 4 / area-level variance of the conditional means model.
Random Intercept Models
Table 1 shows summary statistics for all variables included in the models. Table 2 presents the random intercept estimates for the four outcome variables. The ICC shows the total variation in outcome variables across catchment areas averages between about 5 percent and 9 percent. The remaining 95-91 percent of variance lies within catchment areas.
Table 1
Table 1
SUMMARY STATISTICS
Table 2
Table 2
Random Intercept Variance Estimates Four Explanatory Models and all Dependent Variables
The statistically significant z-scores for all five intercept-only models indicate that PA, BMI and prevalence of obesity vary across catchment areas. The PRE results show that, in most instances, the individual level characteristics account for a higher percent of the variance in the outcome variables across communities than do the environmental variables with the exception of prevalence of obesity. For example, results show that 0.414 or 41.4 percent of the variance in vigorous exercise across communities can be accounted for by adding the individual level variables to the model.
Physical Activity Outcome Measures
Results in Table 3 show physical disorder was significantly negatively associated with students frequently participating in sports. Schools with higher proportions of students who felt unsafe going to or from school were significantly negatively associated with students engaging in all three PA outcome measures. We tested for interactions by gender and found lower perceptions of safety had a negative impact that was much greater for females. The effect was approximately 35 percent greater for vigorous exercise and PA participation and 4 times as great for sports participation. The presence of one additional PA outlet was marginally significantly positively associated with students engaging in frequent vigorous exercise and significantly positively associated with students frequently participating in sports. Neighborhood compactness had a negative significant impact on the likelihood that students frequently participate in sports and traffic danger was marginally significantly negatively associated with student PA participation.
Table 3
Table 3
Results of the Association between Environmental and Physical Activity Measures
Weight-related Outcome Measures
Results in Table 4 show that physical disorder was associated with both higher BMI and the probability of being obese. Neighborhood compactness had a negative and significant impact on the prevalence of obesity and BMI. We conducted sensitivity analyses of the outdoor PA settings scale to determine if certain settings were independently associated with PA and weight. Results showed that the presence of bike paths had a negative significant association with the likelihood of being obese and a negative significant association with BMI.
Table 4
Table 4
Results of the Association between Environmental and Weight Measures
Predicted Probability Models
To explore the relative magnitude of community-level influences on our outcomes, we calculated the probability of changes in PA and BMI-related measures using the coefficients in our models and varying the value of selected predictors one at a time to either the low or high end of the range while holding all other independent variables at their mean. The full set of results is presented in Tables 3 and and4.4. Results of the predicted probability models suggest changes in the built environment could have the greatest impact on decreasing adolescent obesity, i.e., the youth most at risk.
Consistent with existing evidence [10, 1215] we found that increased local area PA outlets are associated with higher levels of PA. However, parents who are active, and in turn encourage or influence their children to be more active, may choose to live in neighborhoods that have more PA outlets and should be examined further in future research.
Contrary to existing evidence, which also controlled for neighborhood income/SES [7,9] we found no association between the presence of parks and sports fields and increased PA or reduced weight. Studies that found an association between outdoor PA settings and increased PA [711] targeted younger age groups. These types of settings may be less important for PA in older youth. Recent research [27] found that park users are primarily children and adults, and that less than 20 percent are adolescents, with more males than females utilizing parks. By simultaneously accounting for multiple measures of the built environment, we are able to assess which environmental measures may have the greatest impact on PA behavior, and may also explain why our findings are inconsistent with previous research. The outdoor PA settings scale is limited in the amount of information it captures. Information on students’ proximity to these settings or on the features of these settings—such as number, type, and condition of playing fields—may have a greater impact on PA than presence alone. This scale is an aggregate measure of these settings, which may bias the estimated effect of our models towards the null making our results more conservative.
Lower levels of neighborhood safety were associated with decreased PA, higher prevalence of obesity and higher BMI. Results showed that perceptions of feeling safe going to and from school are associated with PA. Although our perceived safety measure does not specifically ask about PA, it provides a general indicator of adolescent perceived safety. Previous research [5, 15, 2830] found no association between perceived safety and youth PA. Only one of these studies [15] asked youth specifically about neighborhood safety; this study was conducted in Portugal and the findings may be less relevant for our sample population. Results of interaction effects by gender suggest that perceptions of feeling unsafe going to and from school was associated with greater decreased PA among females across all three outcome variables. This suggests perceived safety is more important for females and may also help to explain why females engage in more indoor physical activities.
We found an association between the level of physical disorder present neighborhoods and decreased sports participation, increased prevalence of obesity and higher BMI. The effects of the physical disorder scale across our outcome variables were somewhat attenuated when the neighborhood median household income measure was added to the models (results not shown). These results imply that some of the correlation between physical disorder and PA and obesity can be attributed to greater presence of these conditions in low-income neighborhoods.
Consistent with previous research [6] we found more compact neighborhoods were associated with lower prevalence of obesity and lower levels of BMI. Neighborhoods characterized as being walkable, or more compact, have also been shown to be associated with adolescent PA [19]. This, coupled with our finding that the presence of bike paths is associated with lower prevalence of obesity and BMI, supports the idea that neighborhoods that are more walkable and bikeable help to increase PA and lower weight. These findings suggest that policies designed to promote more walkable and bikeable communities may have a positive health impact on youth.
More compact neighborhoods were also associated with decreased sports participation. Recent evidence [31] suggests that people living in more compact neighborhoods may substitute different forms of PA (i.e., walking for utilitarian purposes) based on their location, whereas people living in less compact areas may participate more in sports activities. Our finding that students have higher levels of sports participation in less compact neighborhoods may be related to this substitution effect. Alternatively, this may be a reflection that PA venues for youth are different than those for adults. Although our outdoor PA settings scale picks up the presence of sports fields, it does not capture information on the number of venues available in relation to population or size of the area, nor does it capture what types of sports fields or how many were present in the communities. More sprawling areas would have more space for sports fields, which may be more important PA outlets for 8th and 10th graders than walkable compact neighborhoods. There may also be more school and community-based sports programs offered or accessible (i.e., affordable) to youth in more sprawling (i.e., suburban/rural) than urban areas. However, utilitarian walking, the kind that occurs frequently in compact neighborhoods, was not measured in this study. These findings should be investigated further in future research.
Contrary to prior research [32] we found no relationship between traffic danger and the PA measures or weight outcomes. This measure may be more important for leisure and utilitarian walking, and biking to reach PA-related facilities and other neighborhood destinations, but less important for other forms of PA.
Consistent with previous research [33] the results of our random intercept models show the majority of the variance across our outcome measures lies within and not between communities with the environmental measures accounting for 1–2 percent of the total variance in PA and weight. The small magnitude of these effects may in part be due to some of the environmental measures being limited in the amount of information they capture. Again this suggests additional research with more refined environmental measures is needed. Further, community-level variables in general cannot explain much variance in youth outcomes when much of the variance in the outcomes is within communities.
This study was subject to several limitations. First, we used cross-sectional data and cannot make direct causal inferences about whether these environmental measures directly influenced changes in PA behavior and obesity. Second, we captured information on the presence of outdoor PA settings; we did not capture information on how many settings were present or their relative condition. In future studies it would be useful to have information on proximity to these areas in relation to where students live, their accessibility (free vs. paid), as well as a rating system on the availability and condition of sports fields, playground equipment, etc. Third, the PA outlet variable was subject to measurement error due to potential inaccuracies in the commercial outlet density data. Finally, the PA and weight measures were self-reported, which are subject to error and bias. Fourth, we were unable to include a mixed land use measure in our local compactness index. However, previous research [34] indicates that measures of urban form are correlated and higher residential density areas frequently have more mixed land uses and greater street connectivity. Our local compactness index, which includes measures of residential density and street connectivity should serve as a proxy for areas with mixed land use.
Despite these limitations, results from this study revealed important findings. A strength of this study was its ability to simultaneously account for multiple dimensions of the built environment, as well as examine socioeconomic and demographic influences across a range of settings. To our knowledge this is the first study to attempt to understand how varying aspects of a community are associated with youth health behaviors and outcomes using nationally representative data collected directly for communities where the youth live.
Our results show different measures of the environment are associated with PA and weight. We found median household neighborhood income was associated with weight, but not with any of the three PA measures. This suggests that there may be different causal pathways for PA and obesity. The results of the random intercept and the predicted probability models suggest a strong association between the environment and obesity; that changes to the environment could benefit those youth most at risk for future chronic health conditions caused by obesity. It is estimated, with current population growth forecasts, we can expect up to two-thirds of all residential and non-residential buildings will either be replaced or built over the next 40 years [35]. Given this and the growing evidence connecting the built environment to both PA and weight it is important to further explore these associations to help inform and shape future development patterns and land use policies to create neighborhoods that will aid in not only increasing PA behavior, but also help to reduce the prevalence of obesity among youth.
Acknowledgements
Funding for this research was provided by the National Institute on Child Health and Human Development (NICHD) and The Robert Wood Johnson Foundation (RWJF). Monitoring the Future is supported by the National Institute on Drug Abuse. Views expressed are those of the authors and do not necessarily reflect the views of the sponsors, UIC or the University of Michigan. We thank Bridging the Gap colleagues for input into instrument design, Jaana Myllyluoma and colleagues at Battelle Centers for Public Health Research and Evaluation for data collection, and Deborah Kloska and Michael Berbaum for assistance with data analysis.
Footnotes
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 citable 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.
1. Johnston LD, O’Malley PM. Youth, Education, and Society occasional paper #3. Ann Arbor, Institute for Social Research; 2003. Obesity among American adolescents: tracking the problem and searching for causes. http://yesresearch.org/pubs.html.
2. Ogden CL, Carroll MD, Curtin LR, McDowell MA, Tabak CJ, Flegal KM. Prevalence of overweight and obesity in the United States, 1999–2004. JAMA. 2006;295(13):1549–1555. [PubMed]
3. Ogden CL, Carroll MD, Flegal KM. High Body Mass Index for Age Among US Children and Adolescents, 2003–2006. JAMA. 2008;299(20):2401–2405. [PubMed]
4. Papas MA, Alberg AJ, Ewing R, Helzlsouer KJ, Gary TL, Klassen AC. The built environment and obesity. Epidemiologic Reviews. 2007;29(1):144–159. [PubMed]
5. Davison K, Lawson C. Do attributes in the physical environment influence children's physical activity? A review of the literature. International Journal of Behavioral Nutrition and Physical Activity. 2006;3(1):19. [PMC free article] [PubMed]
6. Ewing R, Brownson R, Berrigan D. Relationship between urban sprawl and weight of U.S. youth. American Journal of Preventive Medicine. 2006;31(6):464–474. [PMC free article] [PubMed]
7. Cohen DA, Ashwood JS, Scott MM, Overton A, Evenson KR, Staten LK, et al. Public parks and physical activity among adolescent girls. Pediatrics. 2006;118(5):e1381–e1389. [PMC free article] [PubMed]
8. Brodersen NH, Steptoe A, Williamson S, Wardle J. Sociodemographic, developmental, environmental, and psychological correlates of physical activity and sedentary behavior at age 11 to 12. Ann Behav Med. 2005;29(1):2–11. [PubMed]
9. Gomez JE, Johnson BA, Selva M, Sallis JF. Violent crime and outdoor physical activity among inner-city youth. Preventive Medicine. 2004;39(5):876–881. [PubMed]
10. Timperio A, Crawford D, Telford A, Salmon J. Perceptions about the local neighborhood and walking and cycling among children. Preventive Medicine. 2004;38(1):39–47. [PubMed]
11. Sallis JF, Nader PR, Broyles SL, Berry CC. Correlates of physical activity at home in Mexican-American and Anglo-American preschool children. Health Psychology. 1993;12(5):390–398. [PubMed]
12. Powell LM, Chaloupka FJ, Slater SJ, Johnston LD, O’Malley PO. The Availability of Local Area Commercial Physical Activity-related Facilities and Physical Activity among Adolescents. American Journal of Preventive Medicine. 2007;33(4S):S292–S300. [PubMed]
13. Gordon-Larsen P, McMurray RG, Popkin BM. Determinants of adolescent physical activity and inactivity patterns. Pediatrics. 2000;105(6):E83. [PubMed]
14. 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. Journal of Physical Activity and Health. 2006;3 Suppl 1:S118–S128.
15. Mota J, Almeida M, Santos P, Ribeiro JC. Perceived neighborhood environments and physical activity in adolescents. Preventive Medicine. 2005;41(5–6):834–836. [PubMed]
16. Moore BJ, Glick N, Romanowski B, Quinley H. Neighborhood safety, child care, and high costs of fruit and vegetables as barriers to increased activity and healthy eating and linked to overweight and income. FASEB Journal. 1996;10:A562.
17. Powell LM, Slater SJ, Chaloupka FJ, Harper D. Availability of physical activity-related facilities and neighborhood demographic and socioeconomic characteristics: a national study. American Journal of Public Health. 2007;96(9):1676–1680. [PubMed]
18. Holt NL, Spence JC, Sehn ZL, Cutumisu N. Neighborhood and developmental differences in children’s perceptions of opportunities for play and physical activity. Health and Place. 2008;14:2–14. [PubMed]
19. Carver A, Timperio AF, Crawford DA. Playing it safe: the influence of neighbourhood safety on children's physical activity. A review. Health and Place. 2008;14(2):217–227. [PubMed]
20. Johnston LD, O'Malley PM, Bachman JG, Schulenberg JE. Monitoring the Future national survey results on drug use, 1975–2007. Volume I: Secondary school students. Bethesda, MD: National Institute on Drug Abuse; 2008. NIH Publication No. 08-6418A.
21. Kuczmarski MF, Kuczmarski RJ, Najjar M. Effects of Age on Validity of Self-Reported Height, Weight, and Body Mass Index: Findings From the Third National Health and Nutrition Examination Survey, 1988–1994. Journal of the American Dietetic Association. 2001;101(1):28–34. [PubMed]
22. Dun and Bradstreet. The DUNSright Quality Process: The Power Behind Quality Information. Waltham, Mass: Dun and Bradstreet; 2005.
23. Forsyth A, et al. Environment and Physical Activity: GIS Protocols Version 4.0. University of Minnesota, Metropolitan Design Center; 2006.
24. Ewing R, Schmid T, Killingsworth R, Zlot A, Raudenbush S. Relationship between urban sprawl and physical activity, obesity and morbidity. American Journal of Health Promotion. 2003;18(1):47–57. [PubMed]
25. U.S. Census Bureau. Census 2000 Summary File 3 Technical Documentation. 2002.
26. Raudenbush S, Byrk A. Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, CA: Sage Publications; 2002.
27. Cohen DA, McKenzie TL, Sehgal A, Williamson S, Golinelli D, Lurie N. Contribution of public parks to physical activity. American Journal of Public Health. 2007;97(3):509–514. [PubMed]
28. Burdette HL, Whitaker RC. A national study of neighborhood safety, outdoor play, television viewing and obesity in preschool children. Pediatrics. 2005;116(3):657–662. [PubMed]
29. Adkins S, Sherwood NE, Story M, Davis M. Physical activity among African-American Girls: The role of parents and the home environment. Obesity Research. 2004;12 supplement:38S–45S. [PubMed]
30. Sallis JF, Taylor WC, Dowda M, Freedson PS, Pate RR. Correlates of vigorous physical activity for children in grades 1 through 12: Comparing parent-reported and objectively measured physical activity. Pediatric Exercise Science. 2002;14:30–44.
31. Rodriguez D, Khattak AJ, Evenson KR. Can New Urbanism encourage physical activity? Comparing a New Urbanist neighborhood with conventional suburbs. Journal of the American Planning Association. 2006;72:43–56.
32. Carver A, Timperio AF, Crawford DA. Neighborhood road environments and physical activity among youth: The CLAN study. Journal of Urban Health. 2008;85(4):532–544. [PMC free article] [PubMed]
33. O’Malley PM, Johnston LD, Delva J, Bachman JG, Schulenberg JE. Variation in obesity among American secondary school students by school and school characteristics. American Journal of Preventive Medicine. 2007;33(4S):187–194. [PubMed]
34. Frank LD, Schmid TL, Sallis JF, Chapman J, Saelens BE. Linking objectively measured physical activity with objectively measured urban form: Findings from SMARTRAQ. American Journal of Preventive Medicine. 2005;28(2S2):117–125. [PubMed]
35. Nelson AC. The reconstruction of America. The US Environmental Protection Agency. 2008. http://www.epa.gov/aging/resources/presentations/2008_1028_nelson_reconstruction.pdf.