Article| Volume 23, ISSUE 2, SUPPLEMENT 1, 74-79, August 01, 2002

# The association between urban form and physical activity in U.S. adults

## Abstract

Background: Physical inactivity is associated with multiple adverse health outcomes. Results from the transportation literature suggest that aspects of the urban environment may influence walking for transportation. In this paper we examine the association between a proxy measure of the urban environment and walking behavior.
Methods: We analyzed the association between home age and walking behavior in U.S. adults using data from the Third National Health and Nutrition Examination Survey. Logistic regression was used to estimate odds ratios and 95% confidence intervals and to control for the effects of gender, race/ethnicity, age, education level, household income, and activity limitations.
Results: Adults who lived in homes built before 1946 and from 1946 to 1973 were significantly more likely to walk 1+ miles ≥20 times per month than those who lived in homes built after 1973. This association was present among people living in urban and suburban counties, but absent among those living in rural counties. The association was also found in models that controlled for gender, race/ethnicity, age, education, income, and any health-related activity limitation. Other forms of leisure-time physical activity were not independently associated with home age.
Conclusions: These results support the hypothesis that environmental variables influence walking frequency and suggest that home age may be a useful proxy for features of the urban environment that influence physical activity in the form of walking. Such proxy measures could facilitate testing ecologic models of health behavior using survey data.

## Introduction

Ecologic models of behavior suggest that factors at a variety of levels, from the individual to the community, influence the prevalence of health-related behaviors.
• Moos R.H
The ecologic approach has been widely discussed in the context of physical activity behavior because of the obvious dependence of physical activity on both individual propensity to be active and environmental features that facilitate physical activity.
• Sherwood N.E
• Jeffery R.W
The behavioral determinants of exercise implications for physical activity interventions.
Examples of such features include sidewalks or paths, safe and desirable destinations for walking and cycling, recreational facilities, and a suitable climate.
Health promotion initiatives have already adopted an ecologic model, particularly in the area of physical activity.
• Sallis J.F
• Owen N
Ecological models.
For example, the Centers for Disease Control and Prevention

Centers for Disease Control and Prevention. Active community environments. 2002. Available at: www.cdc.gov/nccdphp/dnpa/aces.htm. Accessed June 1, 2002.

has been promoting the “Active Community Environments Initiative” to enhance activity by modifying the physical environment, but more research is needed to refine our assumptions about ecologic models.
• Sallis J.F
• Owen N
Ecological models.
There is an extensive literature in the transportation research community concerning environmental effects on walking, cycling, and motorized transport use. These studies conceptualize the environment as “urban form” and have focused on transportation systems, such as transit availability and street characteristics, and on land development patterns, such as density and land use mix.
• Frank L
• Pivo G
Impacts of mixed use and density on utilization of three modes of travel single-occupant vehicle, transit, and walking.
,
• Berman M
The transportation effects of neo-traditional development.
Transportation systems and land development are thought to directly influence transportation choices and a number of studies have linked transportation systems and land-use patterns to choices among different transport modes, including walking.
• Ewing R
• Haliyur P
• Page G.W
Getting around a traditional city, a suburban planned unit development, and everything in between.
,
• Kitamura R
• Mokhtarian P.L
• Laidet L
A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area.
In an analysis of data from the 1995 National Personal Transportation Survey, Ross and Dunning
• Ross C.L
• Dunning A.E
showed that the percentage of trips using bicycling or walking as the transport mode was 3.3% in the lowest-density block groups (0 to 99 housing units per square mile), compared to 14.9% in the highest-density block groups (≥3000 units per square mile).
In addition to transportation systems and general patterns of land development, the concept of urban form encompasses building design, building orientation toward the street and other buildings, and detailed aspects of the distribution of homes, workplaces, and other institutions, such as schools, stores, and restaurants.
• Handy S
Urban form and pedestrian choices.
,
• Handy S
Understanding the link between urban form and non-work travel behavior.
The complex nature of urban form makes it difficult and costly to develop comprehensive measures of urban form that may be related to physical activity, and there are significant disagreements over how the existing measures are related to each other and to behavior.
• Berman M
The transportation effects of neo-traditional development.
,
• Kitamura R
• Mokhtarian P.L
• Laidet L
A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area.
,
• Handy S
Urban form and pedestrian choices.
,
• Handy S
Understanding the link between urban form and non-work travel behavior.
,
• Cervero R
Land use mixing and suburban mobility.
,
• Boarnet M
• Crane R
The influence of land use on travel behavior specification and estimation strategies.
Understanding associations between urban form and transportation choices influencing physical activity levels is important for public health because of the possibility that planning decisions could influence physical activity and therefore health.
Home age is a candidate measure of urban form that is worthy of exploration for several reasons. First, home age is associated with density, street design, and building characteristics. Neighborhoods containing older homes in urban areas are more likely to have sidewalks, have denser interconnected networks of streets, and often display a mix of business and residential uses.
• Handy S
Urban form and pedestrian choices.
,
• Handy S
Understanding the link between urban form and non-work travel behavior.
Second, ecologic studies suggest that there may be an association between home age and walking behavior; people living in census tracts with higher-mean home ages walk more than those living in tracts with lower-mean home ages.
• Friedman B
• Gordon S
• Peers J
Effect of neotraditional neighborhood design on travel characteristics.
,
• Douglas I
Lastly, some existing health, transportation, and housing surveys contain data on home age. In this paper, we test the hypothesis that urban form influences levels of physical activity with an analysis of the association between home age, a proxy measure of urban form, and walking behavior in U.S. adults.

## Methods

We analyzed data from the Third National Health and Nutrition Examination Survey (NHANES III). This survey is a nationally representative sample of the U.S. population with a stratified multistage probability design and oversampling of African Americans and Mexican Americans.
National Center for Health Statistics
NCHS plan and operation of the Third National Health and Nutrition Examination Survey, 1988–1994.
A total of 17,030 adults aged ≥20 years responded to the household adult and family survey questions. Our analyses included adults aged ≥20 years who responded to questions concerning all of the behavioral and demographic variables (n = 14,827).
NHANES III survey data included in this study were a measure of walking behavior, measures of the frequency of diverse forms of leisure-time physical activity, home age, rural versus urban/suburban locale, region of the country, demographic characteristics for each respondent, and any health-related activity limitation. Text questions in the survey are available on the web (http://www.cdc.gov/nchs/nhanes.htm). Walking behavior was characterized by responses to the question, “In the past month, how often did you walk a mile or more at a time without stopping?” Note that no attempt was made in the survey to determine the purpose or venue of these walks. This variable had a strongly skewed distribution, with 8334 individuals reporting a frequency of zero times per month. We chose to present the walking data in two ways: (1) categorized as 0 per month, 1 to 19 per month, and ≥20 per month; and (2) categorized as 0 to 19 per month and ≥20 per month. We use the first division for a descriptive analysis of the data and the second bivariate division for logistic regressions. The highest category corresponds to approximately five episodes of walking (as defined in the question) per week.
• Pate R.R
• Pratt M
• Blair S.N
• et al.
Physical activity and public health a recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine.
The measures of leisure-time physical activity all involved questions phrased as, “In the past month did you…?” The eight specified activities were “jog or run,” “ride a bicycle or an exercise bicycle,” “swim,” “do aerobics or aerobic dancing,” “do other dancing,” “do calisthenics or exercises,” “garden or do yard work,” and “lift weights.” If the response was yes, then a subsequent question asked how often this activity was performed. A series of open-ended questions with the same format were also included to allow inclusion of activities not specified. We followed the survey design and treated bicycling as a leisure-time activity rather than a combined estimate of transport and leisure-time activity.
Home age was selected as a proxy measure of environmental factors influencing physical activity. Home age was characterized from responses to the question, “When was this house/structure originally built? Before 1946, 1946–1973, or 1974 to present.” A rationale for these cut-points is not supplied in the NHANES III documentation. In the entire survey, 1505 people responded “don’t know” and 718 responses were classified as “blank but applicable” because of residence in mobile homes. Urban/rural status was evaluated based on categories that were collapsed by the National Center for Health Statistics from the U.S. Department of Agriculture’s rural-urban codes to avoid identification of sampled counties. Urban residences fell in central or fringe counties of metropolitan areas with populations of ≥1 million. Rural areas were all other locations. Lastly, we addressed regional variation in the United States. The data include a code for census region: Northeast, Midwest, South, or West. Regional estimates do not include data for all states in the region, so they are not representative.
National Center for Health Statistics
NCHS plan and operation of the Third National Health and Nutrition Examination Survey, 1988–1994.
Five demographic variables were evaluated: gender, race/ethnicity, age, education level, household income, and a measure of limits to physical activity. Race/ethnicity was reported as non-Hispanic white, non-Hispanic black, Mexican American, and other. Age was categorized into three groups: 20 to 39 years, 40 to 59 years, and ≥60 years. We included measures of education and income in order to account for effects of socioeconomic status. Education was categorized into three categories: less than high school, high school, and any college. Income was divided into two groups: <$20,000 per year and ≥$20,000 per year. More detailed questions were asked about income, but their use resulted in a substantial increase in the rate of nonresponse. Lastly, to account for potential confounding by physical activity limitations, we categorized people as having some limitation if they responded yes to any of three questions. These questions asked if the subject was limited in any way in work, housework, or any activity because of an impairment or health problem.
To account for the complex survey design used in NHANES III, data were analyzed with SUDAAN
• Shah B.B
• Barnwell G
• Bieler G.S
following recommendations discussed in Korn and Graubard.
• Korn E.L
• Graubard B.I
We report frequencies calculated using sample weights associated with the survey design to account for variable selection probabilities. To quantify the relationship between walking behavior and home age, we used logistic regression after dividing the data into urban and rural strata and categorizing walking frequency into two groups: 0 to 19 per month (reference) and ≥20 per month. This makes the odds ratios (ORs) easier to interpret than when three categories are used.
• Korn E.L
• Graubard B.I
(There was no evidence for an association between home age and intermediate walking frequencies.) Home age was not expected to be a proxy variable for environmental features that influence walking in rural areas.
We also examined the association between home age and leisure-time physical activity that did not include walking. If home age is a proxy measure of aspects of urban form that are specifically associated with pedestrian activity, then we hypothesized that other forms of exercise would not be associated with home age in urban or rural areas. To explore potential confounding by demographic variables and any health-related activity limitation, we compared prevalence ORs obtained from models relating walking behavior and home age with and without the inclusion of potential confounders. Lastly, we performed exploratory analyses on our data stratified by region.

## Results

The distributions of walking behavior and home age stratified by several demographic variables are summarized in Table 1. Logistic regression indicated that there was a moderate association between gender and walking frequency (p = 0.062) and strong associations between walking and the remaining variables (p < 0.0001). Men were slightly more likely to walk ≥20 times per month than women. Non-Hispanic whites were more likely than other race/ethnic groups to report some walking and older individuals were most likely to report no walking. High school education and less-than high school education were associated with lack of walking. Home age was not associated with gender (p = 0.41) but it was associated with the remaining demographic variables (p < 0.001). Most apparent among these associations were that older adults were more likely to live in older homes and that lower income and lower levels of education were associated with residence in older structures.
Table 1Demographic and behavioral characteristics of U.S. adults categorized by home age and walking behavior from 1988–1994 NHANES III
CharacteristicTotal (%)Walking frequency (%)When residence built (%)
0/Month1–19/Month≥20/Month≥19741946–1973<1946
n14,8278,3344,5481,9453,6227,3663,839
Weighted %48.738.313.132.843.224.0
Gender
Male48.048.237.714.133.242.624.2
Female52.049.238.812.132.343.823.9
Race/ethnicity
Non-Hispanic white77.747.140.512.433.441.125.5
Non-Hispanic black10.151.732.316.026.752.121.2
Mexican American4.859.228.612.137.048.614.4
Other7.354.628.916.531.249.719.1
Age
20–3945.544.443.911.638.638.722.7
40–5931.447.638.215.034.044.921.1
>6023.158.527.313.519.749.830.5
Education level
<High school23.462.624.512.923.747.029.3
High school33.751.037.012.032.442.625.0
Any college42.939.246.814.038.041.620.4
Household income
<$20,00031.754.332.113.622.645.631.8 ≥$20,00068.346.041.112.837.542.120.4
Activity limitations
Yes15.761.927.011.128.243.328.4
No84.346.240.413.333.643.223.2
NHANES III, Third National Health and Nutrition Examination Survey.
Residents of homes built before 1974 in urban or suburban areas were more likely than residents of newer homes to walk ≥20 times per month (Table 2). This relationship was not seen in rural counties. Results of logistic regressions modeling these associations are shown in Table 3. In urban and suburban areas, adult residents of homes built before 1946 and between 1946 and 1973 were significantly more likely to be in the higher walking category. This result was not affected by inclusion of gender, race/ethnicity, age, education, income, or health-related activity limitation in the model. There was no evidence for an association between home age and other forms of leisure-time physical activity in urban/suburban areas (Table 4). In rural areas, residents of homes built between 1946 and 1973 were less likely to engage in non-walking leisure time than residents of homes built after 1973. However, this association was not present after adjustment for demographic variables (Table 4).
Table 2Distribution of walking frequency and home age in urban/suburban and rural areas for U.S. adults
Walking frequencyTotalWhen residence built
≥19741946–1973<1946
Urban and suburban areas, % (SE)
0/Month45.2 (1.3)47.6 (2.1)44.3 (1.5)43.8 (2.9)
1–19/Month40.1 (1.0)40.8 (2.2)39.6 (1.4)40.2 (1.9)
≥20/Month14.6 (0.8)11.5 (1.0)16.1 (0.9)16.0 (2.1)
Rural counties, % (SE)
0/Month52.0 (1.6)48.2 (2.6)52.8 (1.9)55.7 (2.4)
1–19/Month36.4 (1.5)40.6 (2.0)35.9 (1.9)31.9 (2.0)
≥20/Month11.5 (0.7)11.2 (1.2)11.3 (1.0)12.4 (1.4)
SE, standard error.
Table 3Odds ratios for walking frequencies (≥20 times per month versus <20 times per month) according to home age in U.S. adults in urban/suburban and rural areas
ModelWhen residence built
≥1974 (Reference)1946–1973<1946
Urban and suburban areas
Prevalence, % (SE)
0–19/Month (reference)28.1 (2.6)38.2 (2.3)19.1 (2.6)
≥20/Month3.7 (0.6)7.3 (0.7)3.7 (0.4)
Crude OR (95% CI)1.01.44
Boldfaced ORs have confidence intervals that do not include 1.0.
(1.16–1.79)
1.44 (1.01–2.05)
Adjusted ORs are obtained from regression models including gender, ethnicity, age, education level, income, and health-related activity limitation.
OR (95% CI)
1.01.36 (1.13–1.65)1.43 (1.03–1.98)
Rural counties
Prevalence %, (SE)
0–19/Month (reference)30.0 (2.0)36.3 (1.2)22.1 (2.2)
≥20/Month3.8 (0.5)4.6 (0.4)3.1 (0.4)
Crude OR (95% CI)1.01.00 (0.76–1.34)1.11 (0.75–1.65)
Adjusted OR (95% CI)1.00.93 (0.65–1.33)1.13 (0.74–1.73)
CI, confidence interval; OR, odds ratio.
a Boldfaced ORs have confidence intervals that do not include 1.0.
b Adjusted ORs are obtained from regression models including gender, ethnicity, age, education level, income, and health-related activity limitation.
Table 4Odds ratios for the association between nonwalking, leisure-time physical activity (≥20 times per month versus <20 times per month) and home age in U.S. adults in urban/suburban and rural areas
ModelWhen residence built
≥1974 (Reference)1946–1973<1946
Urban and suburban areas
Prevalence % (SE)
0–19/Month (reference)22.4 (2.7)31.0 (2.4)16.0 (1.3)
≥20/Month9.3 (0.9)14.5 (1.1)6.8 (0.7)
Crude OR (95% CI)1.01.13 (0.87–1.48)1.03 (0.79–1.34)
Adjusted odds ratios are obtained from regression models including gender, ethnicity, age, education level, income, and health-related activity limitation.
OR (95% CI)
1.01.20 (0.88–1.63)1.15 (0.86–1.52)
Rural counties
Prevalence, % (SE)
0–19/Month (reference)22.8 (1.3)29.3 (1.3)17.1 (1.7)
≥20/Month11.0 (1.3)11.7 (0.9)8.1 (0.8)
Crude OR (95% CI)1.00.83
Confidence interval does not include 1.0.
(0.69–0.99)
0.99 (0.78–1.25)
Adjusted OR (95% CI)1.00.87 (0.72–1.06)1.09 (0.89–1.34)
CI, confidence interval; SE, standard error.
a Adjusted odds ratios are obtained from regression models including gender, ethnicity, age, education level, income, and health-related activity limitation.
b Confidence interval does not include 1.0.
Unadjusted analyses stratified by region gave qualitatively similar results for the Northeast, South, and West but not from the Midwest. In urban/suburban areas in the Northeast, South, and West, individuals residing in homes built before 1974 were more likely to report walking ≥20 times per month, with ORs ranging from 1.1 to 1.8. However, 95% confidence intervals for these results overlapped 1.0 for five of six comparisons. In the Midwest, the odds of walking were near or below one for the two older home categories (<1946, OR=0.8; 1946 to 1973, OR=1.0).
We attempted to determine if missing data biased our results by analyzing data that included all individuals reporting home age and walking behavior (n = 15,196). The results of this analysis showed the same pattern as the above crude analysis. The remaining 2000-plus adults that could not be included in the analysis largely consisted of people who did not know the age of their home: They were evenly distributed among the three walking categories (7% to 10% per group).

## Discussion

Past work on the relationship between urban form and physical activity has explored the effects of population density, site design, and building characteristics on transportation behavior. Many of these studies report that increased density and mixed-use development are associated with more walking and bicycling.
• Frank L
• Pivo G
Impacts of mixed use and density on utilization of three modes of travel single-occupant vehicle, transit, and walking.
,
• Ewing R
• Haliyur P
• Page G.W
Getting around a traditional city, a suburban planned unit development, and everything in between.
,
• Kitamura R
• Mokhtarian P.L
• Laidet L
A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area.
,
• Cervero R
Land use mixing and suburban mobility.
,
• Douglas I
,
• Douglas I
,
• Replogle M
However, there is still considerable debate over the consistency and interpretation of such associations.
• Handy S
Urban form and pedestrian choices.
,
• Handy S
Understanding the link between urban form and non-work travel behavior.
,
• Boarnet M
• Crane R
The influence of land use on travel behavior specification and estimation strategies.
,
• Krizek K.J
A pre-test/post-test strategy for researching neighborhood-scale urban form and travel behavior.
In this article, we extend past work by examining the relationship between leisure-time walking behavior and home age, a potential proxy for aspects of urban form that mediate walking behavior. Our finding that walking ≥1 mile ≥20 times per month is associated with living in an urban/suburban residence built before 1974 in U.S. adults supports the hypothesis that home age is a proxy measure of aspects of urban form that influence one aspect of physical activity.
Our findings are also consistent with results from transportation research indicating that walking is more common in neighborhoods with older homes.
• Friedman B
• Gordon S
• Peers J
Effect of neotraditional neighborhood design on travel characteristics.
,
• Douglas I
The current study extends this prior work in at least four ways. First, it is based on data from a nationally representative sample of U.S. adults rather than a local or convenience sample. Second, it includes data on the home age of individual subjects rather than mean home age of a block group or neighborhood; this may or may not be a strength of the study, given that mean home age could be a better measure of neighborhood characteristics than individual home age. Third, it is based on regression models that control for the effects of demographic variables and activity limitations. Fourth, it explicitly contrasts walking with other forms of physical activity in both urban and rural areas.
We found no evidence of an association between home age and walking behavior in rural environments. This suggests that home age in rural environments is not associated with features of the physical environment that influence walking.
• Ross C.L
• Dunning A.E
Regardless of age, homes in rural environments may be too distant from desirable destinations for walking. In addition, environmental variables (e.g., access to parks and trails) that may not be correlated with home age could be mediators of walking frequency in rural areas. Improved understanding of walking behavior could be obtained with an instrument that assessed leisure-time, work-related, and transportation walking behavior along with environmental variables in urban and rural areas.
• Handy S
Urban form and pedestrian choices.
,
• Handy S
Understanding the link between urban form and non-work travel behavior.
,
• Sallis J.F
• Bauman A
• Pratt M
Environmental and policy interventions to promote physical activity.
Home age was not associated with leisure-time activities such as jogging/running, swimming, weight lifting, or dancing. This is consistent with the hypothesis that home age is a proxy measure of urban form that specifically influences pedestrian activity. It also indicates that ecologic analyses of the environment and physical activity should attempt to include diverse measures of environmental characteristics that could be associated with a broader spectrum of activities.
Preliminary data analysis indicated that the association between home age and walking behavior in urban/suburban areas may be present in the Northeast, Midwest, and West, but not in the South. It is difficult to interpret these results because NHANES III was not designed to provide representative samples within regions, and samples from different regions were often obtained at different times of the year.
National Center for Health Statistics
NCHS plan and operation of the Third National Health and Nutrition Examination Survey, 1988–1994.
Furthermore, leisure-time physical activity is known to vary seasonally
• Pell J.P
• Cobbe S.M
Seasonal variations in coronary heart disease.
and to show regional variation in the magnitude of seasonal effects.
Centers for Disease Control and Prevention
Monthly estimates of leisure-time physical inactivity—United States, 1994.
These factors likely complicate the comparison of local or regional cross-sectional studies of physical activity.
• Matthews C.E
• Freedson P.S
• Hebert J.R
• et al.
Seasonal variation in household, occupational, and leisure time physical activity longitudinal analyses from the seasonal variation of blood cholesterol study.
A recent abstract
• Kirkner G
• Levin S
• Durstine J.L
• Hebert J
• Mayo K
Geographic (urban/rural) variations in the prevalence of physical inactivity.
reports that the relationship between physical inactivity and urbanization varies among the four geographic regions of the United States. Geographic variation in attitudes toward physical activity, climatic variables, and safety could moderate the effects of urban form and walking behavior. Our results and the results of Kirkner et al.
• Kirkner G
• Levin S
• Durstine J.L
• Hebert J
• Mayo K
Geographic (urban/rural) variations in the prevalence of physical inactivity.
highlight the potential influence of regional factors such as climate and culture on associations between the environment and walking behavior, but few data are available to address this issue.
A key weakness of cross-sectional studies of the environment and pedestrian activity is that individuals may select neighborhoods based on their preference for walkable environments. To our knowledge, there is only one published longitudinal study of urban form and transportation mode choice.
• Krizek K.J
A pre-test/post-test strategy for researching neighborhood-scale urban form and travel behavior.
This study reported on transportation mode choice of residents of the Puget Sound area. Transport choice data were based on 2-day trip diaries and neighborhood characteristics were coded using aggregate data at the census block-group level. The study found small increases in the use of non-auto transportation (walking, cycling, and transit) when people moved from more to less auto-dependent neighborhoods and small decreases associated with moves to more dependent neighborhoods. Additional longitudinal studies and clever use of natural experiments, such as planned developments or Rails-to-Trails projects, could help us understand the role of urban form in mediating levels of physical activity.
Our study adds to the modest literature documenting associations between measures of urban form and leisure-time physical activity. Furthermore, direct measurements of urban form are costly. The identification of measures of urban form available in existing and ongoing surveys that include measures of physical activity will facilitate research on the environmental correlates of activity.

## Acknowledgements

DB was supported by a Cancer Prevention Fellowship from the National Cancer Institute while working on this project and gratefully acknowledges the Cooper Institute and its sponsors for organizing the symposium in which this paper was first presented. We are grateful to R. Ballard-Barbash, S. Krebs-Smith, K. Dodd, L. Mâsse, A. King, K. Calfas, and A. Bauman for helpful comments on the manuscript. Further comments by three anonymous reviewers also improved the manuscript.

## References

• Moos R.H
The human context. John Wiley & Sons, New York1976
• Sherwood N.E
• Jeffery R.W
The behavioral determinants of exercise.
Annu Rev Nutr. 2000; 20: 21-44
• Sallis J.F
• Owen N
Ecological models.
in: Glanz K Lewis F.M Rimer B.K Health behavior and health education theory, research and practice. Jossey-Bass, San Francisco1997: 403-424
1. Centers for Disease Control and Prevention. Active community environments. 2002. Available at: www.cdc.gov/nccdphp/dnpa/aces.htm. Accessed June 1, 2002.

• Frank L
• Pivo G
Impacts of mixed use and density on utilization of three modes of travel.
Transportation Res Record. 1994; 1466: 44-52
• Berman M
The transportation effects of neo-traditional development.
J Planning Lit. 1996; 10: 347-363
• Ewing R
• Haliyur P
• Page G.W
Getting around a traditional city, a suburban planned unit development, and everything in between.
Transportation Res Record. 1994; 1466: 53-62
• Kitamura R
• Mokhtarian P.L
• Laidet L
A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area.
Transportation. 1997; 24: 125-158
• Ross C.L
• Dunning A.E
Land use transportation interaction. U.S. Department of Transportation, Washington, DC1997
• Handy S
Urban form and pedestrian choices.
Transportation Res Record. 1996; 1552: 135-144
• Handy S
Understanding the link between urban form and non-work travel behavior.
J Planning Educ Res. 1996; 15: 183-198
• Cervero R
Land use mixing and suburban mobility.
Transportation Q. 1988; 42: 429-446
• Boarnet M
• Crane R
The influence of land use on travel behavior.
Transportation Res Part A. 2001; 35: 823-845
• Friedman B
• Gordon S
• Peers J
Effect of neotraditional neighborhood design on travel characteristics.
Transportation Res Record. 1994; 1466: 63-70
• Douglas I
The pedestrian environment, Vol. 4A. 1000 Friends of Oregon, Portland, OR1993
• National Center for Health Statistics
NCHS plan and operation of the Third National Health and Nutrition Examination Survey, 1988–1994.
Vital Health Stat. 1994; 1
• Pate R.R
• Pratt M
• Blair S.N
• et al.
Physical activity and public health.
JAMA. 1995; 273: 402-407
• Shah B.B
• Barnwell G
• Bieler G.S
SUDAAN user’s manual, release 7.5. Research Triangle Institute, Research Triangle Park, NC1997
• Korn E.L
• Graubard B.I
Analysis of health surveys. John Wiley & Sons, New York1999
• Douglas I
Building orientation. A supplement to the Pedestrian Environment, Vol. 4B. 1000 Friends of Oregon, Portland, OR1994
• Replogle M
Integrating pedestrian and bicycle factors into regional transportation planning models. Environmental Defense Fund, Washington, DC1995
• Krizek K.J
A pre-test/post-test strategy for researching neighborhood-scale urban form and travel behavior.
Transportation Res Record. 2000; 1722: 48-55
• Sallis J.F
• Bauman A
• Pratt M
Environmental and policy interventions to promote physical activity.
Am J Prev Med. 1998; 15: 379-397
• Pell J.P
• Cobbe S.M
Seasonal variations in coronary heart disease.
QJM. 1999; 92: 689-696
• Centers for Disease Control and Prevention
Monthly estimates of leisure-time physical inactivity—United States, 1994.
MMWR Morb Mortal Wkly Rep. 1997; 46: 393-397
• Matthews C.E
• Freedson P.S
• Hebert J.R
• et al.
Seasonal variation in household, occupational, and leisure time physical activity.
Am J Epidemiol. 2001; 153: 172-183
• Kirkner G
• Levin S
• Durstine J.L
• Hebert J
• Mayo K
Geographic (urban/rural) variations in the prevalence of physical inactivity.
Med Sci Sports Exerc. 2001; 33: 238