Copyright American Planning Association Winter 2006| [Headnote] |
| Comparing a New Urbanist Neighborhood with Conventional Suburbs |
| If neighborhood design can support or undermine active lifestyles, then residents of new urbanist neighborhoods can be expected to exhibit higher levels of physical activity than residents of conventional communities. This study compared various measures of physical activity for residents of a new urbanist neighborhood to those for a group of conventional suburban neighborhoods in central North Carolina, finding no statistically significant differences, even after adjusting for individual and household characteristics. However, we did detect differences in where people were physically active. Residents of the new urbanist neighborhood were more likely to be physically active in their neighborhood than were residents of conventional suburbs. This difference was due to their walking more for utilitarian purposes, as distinct from walking for leisure. Despite the limitations of a quasi-experimental research design, our results raise questions regarding new urbanism's ability to raise residents' overall levels of physical activity. |
Although rarely cited explicitly as a benefit, existing evidence from studies in established traditional neighborhoods (reviewed in Saelens, Sallis, & Frank, 2003) suggests that residents of new urbanist neighborhoods with supportive design features should be more active physically than residents of less walkable neighborhoods. Thus, neighborhood design could be important in light of the physical inactivity (Centers for Disease Control and Prevention [CDC], 2003) and increasing obesity in the U.S. (Freedman, Khan, Serdula, Galusaka, & Dietz, 2002).
This study responds to recent calls for combining empirical and theoretical approaches from the planning and the health fields to provide a richer perspective for understanding the relationship between physical activity and neighborhood design (Frank & Engelke, 2001; Handy, Boarnet, Ewing, & Killingsworth, 2002; Sallis, Kraft, & Linton, 2002). We matched a large new urbanist community with conventional suburban neighborhoods to examine physical activity patterns. We collected data using a travel diary adapted from the 2001 National Household Transportation Survey (U.S. Department of Transportation, 2003) and questions on moderate and vigorous physical activity adapted from the 2001 Behavioral Risk Factor Surveillance System (BRFSS) survey (National Center for Chronic Disease Prevention and Health Promotion, 2004). In addition, we examined the degree to which individuals substitute physical activity in one location for physical activity in other locations-a question that has not been tested previously.
The Connection between Neighborhood Design and Physical Activity: Prior Evidence
Travel is a major potential source of physical activity in modern life. Residential location and neighborhood design might influence physical activity through affecting destination and mode choices. Accumulating evidence suggests that people's decisions to walk or cycle are related to certain characteristics of the built environment. Cross-sectional studies have found positive associations between physical activity obtained through travel and the presence of mixed land uses (Cervero, 1996; Cervero & Kockelman, 1997; Moudon, Hess, Snyder, & Stanilov, 1997; Saelens, Sallis, & Frank, 2003), improved street connectivity (Boarnet & Crane, 2001; Boarnet & Sarmiento, 1998; Crane & Crepeau, 1998; Greenwald & Boarnet, 2000; Kitamura, Laidet, & Mokhtarian, 1997), and higher employment and population density at origins and destinations (Cervero, 1996; Cervero & Wu, 1997; Frank & Pivo, 1994; Messenger & Ewing, 1996). Commuting studies have found that a neighborhood's location and connectivity to other activity centers in the region are critical in explaining observed travel patterns, and thus potentially to the physical activity obtained through that travel (Cervero & Kockelman, 1997; Ewing, 1995; Ewing & Cervero, 2001; Kasturi, Sun, & Wilmot, 1998; Saelens, Sallis, & Frank, 2003).
Leisure activities provide another potential source of physical activity. Recent reviews of environmental factors associated with physical activity among adults published in public health journals (Humpel, Owen, & Leslie, 2002; Owen, Humpel, Leslie, Bauman, & Sallis, 2004; Saelens, Sallis, & Frank, 2003) highlight that relatively few consistent relationships between leisure-time physical activity and the built environment have been identified. However, these studies have developed almost independently from the study of physical activity resulting from travel behavior (Handy, 1992; Saelens, Sallis, Black, & Chen, 2003; and Troped, Saunders, Pate, Reininger, & Addy, 2003 are notable exceptions). Of 38 public health studies on physical activity predictors among adults reviewed by Trost et al. (2002), the majority focused on leisure-time activity.
Other studies (reviewed in Ewing & Cervero, 2001; and Saelens, Sallis, & Frank, 2003) have compared travel patterns in neighborhoods with and without environmental characteristics hypothesized to support walking, while matching them on other characteristics such as regional accessibility and residents' income. If these neighborhoods truly differ only on the variables of interest, such a quasiexperimental research design has the advantage of providing a comprehensive comparison of the "package" of built environment elements present in the neighborhoods. In the only study to date that has compared physical activity of residents in two neighborhoods, Saelens, Sallis, Black, and Chen (2003) found that total physical activity measured using accelerometers was over 50% higher in a highly walkable (but not new urbanist) neighborhood as compared to a less walkable neighborhood of similar density. However, self-reported leisure-time activity and self-reported walking for exercise were no different. They also found that more residents of the less walkable neighborhood were overweight, and that moderate rather than vigorous activity accounted for the difference in physical activity.
Despite important work to date, theoretical questions and empirical gaps remain (Badoe & Miller, 2000; Bauman, Sallis, Dzewaltowski, & Owen, 2002; Boarnet & Crane, 2001; Crane, 2000; Ewing & Cervero, 2001; Humpel, Owen, & Leslie, 2002; Saelens, Sallis, & Frank, 2003; Trost et al., 2002). Theoretical critiques have highlighted the lack of a strong behavioral foundation for hypotheses and anticipated relationships. Multilevel models of health behavior have begun to address some of these theoretical gaps (e.g., the socioecologic model of Stokols, 1992, and the social determinants of health and environmental health promotion model of Northridge, Sclar, & Biswas, 2003). However, limitations in research design, measurement, and data collected hamper the comparability and transferability of results, and the existing evidence raises many questions regarding characteristics that seem to be related to physical activity. We need additional evidence regarding the magnitude and causal direction of the relationship between microlevel measures of the built and natural environments and physical activity. One important question for neighborhood-based studies is whether physical activity is greater in neighborhoods that support walking than in conventional suburban neighborhoods. Also, are residents of various types of neighborhoods physically active in the same ways? Directly observing the environments where physical activity takes place could be helpful in clarifying relationships. Likewise, if barriers to physical activity are removed, does total physical activity increase or do individuals substitute physical activity in one location for physical activity in the location with fewer barriers?
Research Design
We followed a quasi-experimental research design (Shadish, Cook, & Campbell, 2002), matching a large new ubanist neighborhood with a group of conventional suburban neighborhoods located in the Chapel Hill-Carrboro area of central North Carolina. The new urbanist neighborhood was a greenfield development built in the late 19905 and early 20005, and features small lot sizes, office and commercial space within walking distance of most residences, a variety of residential options (single-family homes, townhomes, and condominiums), amenities for pedestrians and bicyclists, reduced building setbacks, and rear alleyways for garages and services like garbage collection and mail delivery.
To identify suitable matches to the new urbanist neighborhood, we inventoried all contemporaneous planned unit developments and conventional subdivisions, limiting candidates to Chapel Hill and Carrboro to control experimentally for differences in location, school quality, and other public services. We evaluated neighborhoods based on their proximity to local employment sites, gross tract area, number of housing units, average assessed value of homes, age of the development, transit service availability, regional accessibility, aesthetics, safely, and land use mix (for details on the evaluation, see Khattak et al., 2004). Although no single conventional neighborhood met all the matching criteria, we selected a group of spatially clustered neighborhoods that together provide a suitable match (see Figure 1). Table 1 summarizes key characteristics.
To guide our examination of the relationship between physical activity and neighborhood attributes, we rely on a socioecological conceptual model, portrayed in Figure 2. The arrows in the model show influences on physical activity at various ecological levels. Solid lines represent associations tested in this study, while dashed lines represent hypothesized associations not tested. The model considers the built environment an intermediate factor that, together with macro-, interpersonal-, and individual-level factors, contributes to and also potentially influences physical activity behavior. It emphasizes connections among factors at multiple levels and possibly iterative relationships. Thus, for example, our conceptual model captures how individual preferences for built environments can influence physical activity behaviors through their role in the selection of a residential neighborhood.
 | |
| Figure 1. Site plans for conventional neighborhoods (left) and the new urbanise neighborhood (right). |
A survey instrument distributed by mail to all heads of household in both areas comprised two sections. Household heads filled out the first section, relating to household demographic and physical characteristics (height and weight); travel patterns; attitudes and preferences for the built environment; and physical activity frequency and location (at home, in the neighborhood but not at home, or outside of the neighborhood). Section two of the survey, covering activity by all household members aged 16 and up, was a travel diary based on the 2001 National Household Transportation Survey (NHTS).
Outcome Measures
We assessed physical activity using questions from the BRFSS survey and the NHTS travel diary. For the former, we used the year 2001 BRFSS survey module on moderate and vigorous physical activity (CDC, 2003). We asked heads of households whether they participated, except as part of their work, in "any moderate activity for at least 10 minutes at a time, such as brisk walking, bicycling, vacuuming, gardening, or anything else that causes some increase in breathing or heart rate" in a usual week. We also asked about "any vigorous activity for at least 10 minutes at a time, such as running, aerobics, heavy yard work, or anything else that causes large increases in breathing or heart rate." A "yes" to either question prompted a query about the frequency and total time spent on these activities daily. Others have established the reliability of these questions (Evenson, Eyler, Wilcox, Thompson, & Burke, 2003; Evenson & McGinn, in press).
 | |
| Table 1. Characteristics of the built environment in study neighborhoods. |
From these data, we constructed several measures of physical activity for each participant: (a) total time devoted to moderate and vigorous physical activity; (b) whether they met or did not meet current recommendations for weekly physical activity, defined as being moderately active for at least 30 minutes for 5-7 days a week or vigorously active at least 20 minutes for 3-7 days a week (Pate et al., 1995); and (c) whether, if they did not meet recommended levels of physical activity, they were insufficiently active (defined as somewhat physically active but not enough to meet recommendations) or inactive (participating in less than io minutes of moderate or vigorous physical activity in a usual week).
To explore the possibility that physical activity in other locations substitutes for physical activity at or near home, we asked heads of household to tell us the percentage of moderate and vigorous physical activity time they spent at home, outside the home but in the neighborhood, and outside the neighborhood. With this information, we computed total time devoted to moderate and vigorous physical activity by location. We hypothesized that residents of the new urbanist neighborhood were more likely to be physically active in their neighborhood than residents of the conventional neighborhoods, who we hypothesized to be physically active elsewhere (outside of their neighborhoods or in their homes).
The travel diary responses provided detailed information on physically active travel for every adult household member, including the head of household. To capture the concept of physically active travel, we expanded on the NHTS definition of a trip to include travel "from one place to another" involving "movement of more than 300 feet, including walking for exercise, walking dogs, and riding bikes." For each such trip, we asked participants to record travel time, destination, mode, travel distance, and out-of-pocket costs. We assumed pedestrian or bicycle trips involved moderate or vigorous physical activity, and travel by all other modes involved no moderate or vigorous physical activity. We coded trips that began and ended at home as leisure trips, and other trips as utilitarian trips. That is, utilitarian trips were defined as those derived from the desire to access a destination like purchasing groceries, going to the movies, dropping off a relative somewhere, or visiting a friend. By contrast, leisure trips were defined as trips not involving destinations, and therefore included walking a dog, taking a stroll, or exercising without accessing a destination. We calculated total duration and frequency of all physically active trips, as well as distinguishing physically active utilitarian from leisure trips.
 | |
| Figure 2. Conceptual framework for understanding physical activity behavior indicating relationships tested in the current study. |
Limitations
Our research design and outcome measures have several limitations. First, because our study relies on cross-sectional data, we must interpret any relationships we discover solely as associations, not causative factors. Any relationships between the built environment and physical activity discovered with this research design may be the result of preexisting individual preferences rather than a change caused by the environment. This seems likely, since the conceptual model (shown in Figure 2) suggests that individuals or households with preferences towards certain health behaviors (like physically active utilitarian travel) may choose built environments that support those behaviors. However, though an individual's ability to choose their own residence threatens our ability to establish cause and effect, it could be the mechanism through which neighborhood design affects behavior. Were neighborhoods with new urbanist characteristics in short supply (as suggested by Levine, 2005; and Morrow-Jones, Irwin, & Roe, 2004), those inclined to choose such neighborhoods would be limited to conventional suburbs. Thus, without new urbanist neighborhoods, any physical activity enabled by such neighborhoods, whether by satisfying or modifying pre-existing preferences, would be reduced.
A second limitation is that participants self-reported the duration and location of physical activity, which they may have perceived or remembered incorrectly, or deliberately misreported. Future studies may rely on objective measures of physical activity duration and location (for example using global positioning systems) to determine whether our measures are reliable and valid.
Third, we are limited in our ability to generalize to other regions, although the physical characteristics (median block size, street connectivity, and the presence of nonresidential land uses) of the new urbanist neighborhood and the control area we used are remarkably similar to what Song and Knaap (2004) described as a representative new urbanist development in Portland, Oregon, and its conventional counterpart. Our low response rate further hampers our ability to generalize. Future research will show whether the relationships we found exist elsewhere. Until then, assuming our results do generalize to other communities in the Southeast, they may inform planning in this region of the U.S., which has experienced the greatest increase (67%) in the percentage of the population that is overweight from 1991 to 1998 (Mokdad et al., 1999). Obesity is particularly prevalent in North Carolina (Mokdad et al., 2001), where BRFSS data indicate that a smaller proportion of adults achieve the recommended level of activity than in the rest of U.S. (N.C. Center for Health Statistics, 2005).
Descriptive Statistics
From March through May 2003, we distributed 920 surveys in the new urbanist neighborhood and 891 surveys in the conventional neighborhoods to be filled out by household heads. Each survey also included five travel diaries to be filled by household members aged 16 and older. Our initial response rates were 26.4% (n=243) and 23.6% (n= 210) for the new urbanist and conventional neighborhoods, respectively. We ended up with 393 valid household head survey responses and 370 valid household head travel diaries.1 We had travel diaries for a total of 711 individuals using the data for all household members.2
Table 2 shows demographic characteristics of household heads in each neighborhood, stratified by dwelling type. The conventional neighborhoods had no multifamily units. Comparisons of variables between the two neighborhood types were made using unpaired t-tests assuming unequal variances for continuous variables and Fisher's exact test for dichotomous variables. Statistical significance tests were two-tailed. Residents of multifamily units in the new urbanist neighborhood were more likely to be younger, female, and students; have smaller households; and have fewer automobiles than residents of the conventional neighborhoods. Comparison of residents of single-family homes in the two types of neighborhood also shows these differences. Residents of single-family units in the new urbanist neighborhood are more likely to be female, have fewer automobiles, and have fewer household members than such residents of conventional neighborhoods. These differences may be a source of systematic differences in physical activity.
Results
Table 3 shows that residents of the new urbanist neighborhood living in either multi- or single-family units did not engage in more or less physical activity than residents of the conventional neighborhoods. Differences in the shares of residents meeting recommended levels of physical activity were not statistically significant among the neighborhood types and dwelling types; 50% of those in conventional neighborhoods, 55% of those in single-family units in new urbanist neighborhoods, and 48% of those in multifamily units in new urbanist neighborhoods meet recommended levels. Likewise, we found no significant differences among the shares of residents inactive, insufficiently active, and meeting recommended activity levels in the different neighborhood types.
| Table 2. Sociodemographic characteristics of household heads by neighborhood and dwelling type. |
However, comparisons of where moderate or vigorous physical activity occurs for residents of both types of neighborhood showed that residents of the new urbanist neighborhood spent more time being physically active in their neighborhood than did residents of the conventional neighborhoods. Residents of the new urbanist neighborhood spent less time in moderate or vigorous physical activity inside their homes and outside of their neighborhoods than did residents of the conventional neighborhoods, although this difference is significant only for moderate physical activity among multifamily unit dwellers in their homes. We found comparable differences when classifying persons as inactive, insufficiently active, and meeting recommended activity levels, though these are not shown. Taken together, these results support our hypothesis that differences in the built environment do affect where physical activity occurs.
Table 4 summarizes travel diary data for walking and cycling. The total number of walking and cycling trips per day for single-family household heads was 2.4 times higher in the new urbanist neighborhood than in the conventional neighborhoods (0.92/0.39). The number of recreation/ leisure walking and cycling trips reported in each neighborhood type is quite similar to those reported in two other studies examining highly walkable neighborhoods versus those less suited to walking (Handy, 1992, 1996). However, the total number of these trips (including recreation/leisure and utilitarian trips) is somewhat higher than in most other studies (Friedman, Gordon, & Peers, 1994; Handy & Clifton, 2001; McNally & Kulkarni, 1997) and similar to only one other study reviewed (Kitamura et al., 1997).
| Table 3. Reported hours of physical activity per week among household heads by neighborhood and dwelling type. |
Table 4 also shows that the new urbanist neighborhood dwellers not only made more walking and cycling trips, but also spent almost twice as much total time walking and cycling as residents of the conventional neighborhoods. This result is particularly important for assessing the health consequences of the neighborhood designs, since the durations of physical activity episodes can impact health outcomes (Saelens, Sallis, & Frank, 2003).
Thus far we have compared mean values for the outcome measures across neighborhood types. Since individual-level variables tend to be the strongest and most consistent predictors of physical activity behavior (Giles-Corti & Donovan, 2002; Sallis & Owen, 1999), one potential explanation for our results is that there are systematic differences between residents of the different neighborhood types, causing spurious associations. To account for this, we estimate 12 regression models, one for each outcome measure. We use ordinary least squares (OLS) regression for continuous outcome measures, logistic regression for categorical outcome measures, and negative binomial regression for outcome measures involving count data. Each regression model includes two dummy variables of interest, one indicating residence in single-family dwellings in the new urbanist neighborhood and the other indicating residence in multifamily dwellings in the new urbanist neighborhood. The models also control for other confounders, including gender, age, occupational status (for household heads), vehicles per household, and household size.3
Table 5 summarizes the results for heads of household.4 Though they are not shown here, models including data for all household members display very similar results. At least one of the two variables indicating residence in the new urbanist neighborhood is significant at standard levels of confidence in 7 of the 12 models. Neither is significant in Models 1, 2, 3, 4, and 6. Each model contains the same control variables, including those that were not significant, to ensure comparability. In the OLS and median regression models, the coefficients on the dummy variables indicating residence in particular dwelling types in the new urbanist neighborhood quantify the effect of such residence on the average individual (e.g., 0.6 hours extra of physical activity per week for single-family housing). The logistic regression coefficients on these variables measure the effect of such residence on the log-odds of the dependent variable, and the negative binomial coefficients on these variables measure the effect of such residence on the log of the number of trips.5
 | |
| Table 4. Walking and cycling trips and travel time per day for household heads by neighborhood and dwelling type. |
The first three models confirm what the descriptive statistics suggested; there was no statistically significant difference between the total time spent in moderate or vigorous physical activity by residents of the two neighborhood types, nor in whether or not they meet activity requirements or are categorized as active. Results of Models 4, 5, and 6 demonstrate that residents of the new urbanist neighborhood, regardless of dwelling type, devoted about one hour more per day to physical activity in the neighborhood than did residents of the conventional neighborhoods. The negative signs of the coefficients in Models 4, 5, and 6 suggest that they also allocated less time to physical activity in their homes and outside their neighborhoods than did residents of the conventional neighborhoods, but these differences did not reach statistical significance.
Models 7, 8, and 9 show the effects of neighborhood and dwelling type on the number of bicycle and pedestrian trips, and Models 10, 11, and 12, show the effect of neighborhood and dwelling type on time spent in travel by those two modes. With a few exceptions noted below, the pattern of association summarized in Table 4 remains. Residents of the new urbanist neighborhood living in single-family homes made 2.1 times more pedestrian and bicycle trips (e^sup 0.76^), and diose living in multifamily units made 1.9 times more pedestrian and bicycle trips (e^sup o.62^), than did residents of the conventional neighborhoods. For heads of households living in single-family dwellings, residence in the walkable, new urbanist neighborhood was associated with more utilitarian pedestrian and bicycle trips (Model 9), but the same number of recreational pedestrian and bicycle trips (Model 8), consistent with prior research (Handy, 1992., 1996; Handy & Clifton, 2001; Saelens, Sallis, Black, & Chen, 2003).
One question that emerges is whether residents of the new urbanist neighborhood are making more utilitarian walking and bicycle trips because they are making more trips overall, or because they are reducing trips by other modes. The multimodal travel diary data collected in this study and analyzed elsewhere (see Khattak & Rodriguez, 2005) suggests that the total number of per capita trips, regardless of travel mode, is similar in the two neighborhood types, and that the higher utilitarian walking activity occurs at the expense of travel by car. After controlling for sociodemographics and related variables, we found that households in the new urbanist neighborhood made 1.6 fewer auto trips and traveled 14.7 fewer vehicle miles per day than did those in the conventional neighborhoods.
| Table 5. Associations between neighborhood type and physical activity of household heads, controlling for sociodemographic and household characteristics. |
Finally, the time spent in pedestrian and bicycle travel (Models io, n, and 12) provide a more nuanced picture of physical activity through travel than is found in the existing literature, with heads of households living in the new urbanist neighborhood spending 0.13 more hours each day walking and cycling than do their counterparts living in the conventional neighborhoods. This translates into household heads obtaining 55 more minutes of moderate physical activity per week from utilitarian travel in the neighborhood (or 42 additional minutes per week on average among all household members). As with trip frequency, virtually all the difference in cumulative trip duration results from utilitarian trips.
Discussion and Conclusions
For planners and researchers, this study yields mixed evidence regarding the relationship between the built environment and physical activity. Residents of the neighborhood possessing attributes believed to support physical activity are more physically active in their neighborhood, making approximately twice as many trips and logging 40 to 55 minutes more walking and cycling each week than their counterparts in the conventional suburban neighborhoods. Utilitarian travel, rather than leisure travel, is largely responsible for these differences. Assuming that we captured important design differences by comparing our two neighborhood types, we concur with Saelens, Sallis, Black, and Chen (2003) that neighborhood design is not related to leisure-time physical activity when one controls for individual- and household-level characteristics.
It is also encouraging to note that increased numbers of walking trips came at the expense of automobile trips, consistent with prior evidence (Cervero & Kockelman, 1997; Cervero & Radisch, 1995) that residents of walkable neighborhoods traveled fewer vehicle miles and made more walking trips than residents of conventional suburban neighborhoods. Even as an example of an island "in a sea of freeway-oriented suburbs" (Cervero, 1996), greenfield new urbanist developments like the one studied here enable socially preferable travel behavior by supporting alternative modes of transportation and decreasing vehicle miles traveled, which translate into fewer cars on the roadway.
However, we did not observe differences in physical activity between heads of households in the two neighborhood types using any of three different outcome measures. Thus, although heads of households in the new urbanist neighborhood were more physically active in their neighborhood, they appeared to be less physically active elsewhere. This substitutive behavior, which to our knowledge has not been described in the existing literature, is consistent with the idea that individuals of like socioeconomic status may have similar time budgets for physical activity. Furthermore, our results may help explain other seemingly surprising findings in the literature. For example, a recent study comparing six communities that actively promote use of walking trails to six communities that do no such promotion found that although use of walking trails doubled, total walking activity did not change (Brownson et al., 2004).
Because we studied only one new urbanist neighborhood and obtained only a modest survey response rate, these results are not necessarily generalizable unless they are replicated in other contexts and for populations with different socioeconomic attributes (Ford et al., 1991). In particular, other studies may investigate whether the substitutive behavior we observed is widespread. Furthermore, differences in income may help explain why we find no differences in physical activity by neighborhood type while Saelens, Sallis, Black, and Chen (2003) do identify such differences. The participants in their study resided in tracts with median yearly household income in the low- to mid$40,0005, whereas the median tract income for respondents in our study area is about twice that value.
Our results have additional implications for practicing planners. First, our findings do not support the hypothesis that residents of new urbanist neighborhoods will be more physically active than residents of conventional suburban neighborhoods. second, the results reinforce the view that bringing origins and destinations closer together is related to increased walking and the incorporation of physical exercise into utilitarian travel, though this may substitute for other physical activity. The extent to which this association is causal is a matter of empirical debate, but the fact remains that residents of new urbanist neighborhoods exhibit higher levels of walking activity-whether due to previously held preferences or to the effect of neighborhood design. Third, since leisure-time physical activity increases with income (Pate et al., 1995; U.S. Department of Health and Human Services, 2001), new urbanism's support for physical activity obtained through utilitarian travel may be important for groups with lower incomes than those in the current study. Fourth, the fact that residents of new urbanist neighborhoods spend more time being physically active in their neighborhood indicates a possible explanation for claims of a stronger sense of community and higher neighborhood cohesion documented elsewhere (Kim & Kaplan, 2004; Lund, 2003). Our conceptual model also provokes us to ask whether residents of new urbanist neighborhoods, by spending more time outdoors in their neighborhoods, cause an increase in social cohesion over time, and perhaps as a result reinforce increased physical activity. However, to the extent that we can generalize from our case, we do not find evidence that new urbanism promotes higher levels of total physical activity.
Acknowledgements
Financial support for this study was provided in part by the North Carolina Department of Transportation. We are also very grateful to the NCDOT Research and Analysis Group for their support during the project. The study was conducted under the auspices of the Carolina Transportation Program, 
University of North Carolina at Chapel Hill. We are grateful to Austin Brown, Ben Rasmussen, Steve Wernick, David Anspacher, Jennifer Valentine, and Yinglin Fan for their assistance in completing this study and to the anonymous referees.
| [Footnote] |
| Notes |
| 1. We eliminated records for three household heads for whom we were missing age, gender, or vehicle ownership data, and 52. household heads for whom we were missing BRFSS physical activity data, as well as five household heads reporting physical activity more than five standard deviations from the mean. This resulted in 393 valid household head responses to the survey. Our final sample size for travel diaries from household heads was 370, after dropping 76 observations missing travel diary data, two with travel activity more than five standard deviations from the mean, and five without sociodemographic information. In all cases, results of statistical tests achieved the same levels of significance regardless of whether we included outlier observations. |
| 2. Of these, five observations were missing age information. We imputed the value of age for these five observations and created an indicator variable with a value of 1 if the observation was imputed and o otherwise. When used concurrently with age in a regression equation, the indicator variable captures any bias resulting from the imputation of the age values, while allowing us to work with a larger sample size. |
| 3. We do not control for household income because one of the criteria for matching the two neighborhoods was assessed property values. Thus, we are controlling for personal wealth effects on behavior with the research design. Other variables such as occupational status will capture intraneighborhood variations of relevance. |
| 4. The residuals of the first model of MVPA time were bimodally distributed. To remedy this, we used a median regression estimation routine (also known as least absolute value models; see Bloomfield & Steiger, 1980). |
| 5. Because the negative binomial model is multiplicative, the anti-log of the coefficient measures the factor change in the number of trips. For details, see Long (1997). |
| [Reference] » View reference page with links |
| References |
| Badoe, D., & Miller, E. (2000). Transportation-land use interaction: Empirical findings in North America, and their implications for modeling. Transportation Research Part D, 5(4), 235-263. |
| Bauman, A. E., Sallis, J. F., Dzewaltowski, D. A., & Owen, N. (2002). Toward a better understanding of the influences on physical activity: The role of determinants, correlates, causal variables, mediators, moderators, and confounders. American Journal of Preventive Medicine, 23(2Suppl.), 5-14 |
| Bloomfield, P., & Steiger, W. (1980). Least absolute deviation curvefitting. SIAM Journal of Scientific and Statistical Computing, 1, 290-301. |
| Boarnet, M. G., & Crane, R. (2001). Travel by design: The influence of urban form on travel. New York: Oxford University Press. |
| Boarnet, M. G., & Sarmiento, S. (1998). Can land-use policy really affect travel behavior? A study of the link between non-work travel and land-use characteristics. Urban Studies, 35(7), 1155-1169. |
| Brownson, R. C., Baker, E. A., Boyd, R. L, Caito, N. M., Duggan, K., Housemann, R. A., et al. (2004). A community-based approach to promoting walking in rural areas. American Journal of Preventive Medicine, 27(1), 28-33. |
| Centers for Disease Control and Prevention. (2003). Prevalence of physical activity, including lifestyle activities among adults-United States, looo-iooi. Morbidity and Mortality Weekly Report, 52(31), 764-769. |
| Cervero, R. (1996). Mixed land-uses and commuting: Evidence from the American Housing Survey. Transportation Research A, 30(5), 361-377. |
| Cervero, R., & Kockelman, K. (1997). Travel demand and the 3Ds: Density, diversity and design. Transportation Research D, 2(3), 199-219. |
| Cervero, R., & Radisch, C. (1995). Travel choices in pedestrian versus automobile-oriented neighborhoods. Transport Policy, 5(3), 127-141. |
| Cervero, R., & Wu, K.-L. (1997). Polycentrism, commuting, and residential location in the San Francisco Bay Area. Environment and PlanningA, 29, 865-886. |
| Crane, R. (2000). The influence of urban form on travel: An interpretative review. Journal of Planning Education and Research, 15(1), 3-23. |
| Crane, R., & Crepeau, R. (1998). Does neighborhood design influence travel? A behavioral analysis of travel dairy and GIS data. Transportation Research D, 5(4), 225-238. |
| Evenson, K. R., Eyler, A. A., Wilcox, S., Thompson, J. L., & Burke, J. E. (2003). Test-retest reliability of a questionnaire on physical activity and its correlates among women from diverse racial and ethnic groups. American Journal of Preventive Medicine, 25(3 Suppl. 1), 15-22. |
| Evenson, K. R., & McGinn, A. (in press). Test-retest reliability of adult surveillance measures of physical activity and inactivity. American Journal of Preventive Medicine. |
| Ewing, R. (1995). Beyond density, mode choice, and single-purpose trips. Transportation Quarterly, 49, 15-24. |
| Ewing, R., & Cervero, R. (2001). Travel and the built environment. Transportation Research Record, 1780, 87-114. |
| Ford, E. S., Merritt, R. K., Heath, G. W., Powell, K. E., Washburn, R. A., Kriska, A., et al. (1991). Physical activity behaviors in lower and higher socioeconomic status populations. American Journal ofEpidemi' ology, 133(12), 1246-1256. |
| Frank, L., & Engelke, P. O. (2001). The built environment and human activity patterns: Exploring the impacts of urban form on public health. Journal of Planning Literature, 16(1.), 202-218. |
| Frank, L., & Pivo, G. (1994). Impacts of mixed use and density on utilization of three modes of travel: Single-occupant vehicle, transit, and walking. Transportation Research Record, 1466, 44-52. |
| Freedman, D. S., Khan, L K., Serdula, M. K., Galuska, D. A., & Dietz, W. H. (2002). Trends and correlates of class 3 obesity in the United States from 1990 through 2000. Journal of the American Medical Association, 288(14), 1758-1761. |
| Friedman, B., Gordon, S. P., & Peers, J. B. (1994). Effect of neotraditional neighborhood design on travel characteristics. Transportation Research Record, 1466, 63-70. |
| Giles-Corti, B., & Donovan, R. (2002). Socioeconomic status differences in recreational physical activity levels and real and perceived access to supportive physical environment. Preventive Medicine, 35(6), 601-611. |
| Greenwald, M. J., & Boarnet, M. G. (2000). Built environment as a determinant of walking behavior: Analyzing nonwork pedestrian travel in Portland, Oregon. Transportation Research Record, 1780, 33-42. |
| Handy, S. (1992). Regional versus local accessibility: Neo-traditional development and its implications for non-work travel. Built Environment, 18, 253-267. |
| Handy, S. (1996). Urban form and pedestrian choices: Study of Austin neighborhoods. Transportation Research Record, 1552, 135-144. |
| Handy, S., Boarnet, M., Ewing, R., & Killingsworth, R. (2002). How the built environment affects physical activity-Views from planning. American Journal of Preventive Medicine, 25(28), 64-73. |
| Handy, S., & Clifton, K. J. (2001). Local shopping as a strategy for reducing automobile travel. Transportation, 28, 317-346. |
| Humpel, N., Owen, N., & Leslie, E. (2002). Environmental factors associated with adults' participation in physical activity. American Journal of Preventive Medicine, 22(3), 188-199. |
| Kasturi, T., Sun, X., & Wilmot, C. G. (1998). Household travel, household characteristics, and land use: An empirical study from the 1994 Portland activity-based travel survey. Transportation Research Record, 1617,10-17. |
| Khattak, A., & Rodriguez, D. (2005). Travel behavior in neo-traditional neighborhoods: A case study in USA. Transportation Research Part A, 30(6), 481-500. |
| Khattak, A., Stone, J., Rodriguez, D., Letchworth, W. E., Rasmussen, B. K., & Schroeder, B. J. (2004). Traditional Neighborhood Development Trip Generation Study-Drafifinal report (No. 2003-13). Raleigh: North Carolina Department of Transportation. |
| Kim, J., & Kaplan, R. (2004). Physical and psychological factors in sense of community-New urbanist Kentlands and nearby Orchard Village. Environment and Behavior, 36(3), 313-240. |
| Kitamura, R., Laidet, L., & Mokhtarian, P. (1997). A micro-analysis of land use and travel in five neighborhoods in the San Francisco Bay Area. Transportation, 24, 125-158. |
| Levine, J. C. (2005). Zoned out: Regulation, markets, and choice in transportation and land use. Washington, DC: Resources for the Future Press. Long, S. (1997). Regression models for categorical and limited dependent variables. Thousand Oaks, CA: Sage. |
| Lund, H. (2003). Testing the claims of New Urbanism. Journal of the American Planning Association, 60(4), 414-429. |
| McNally, M. G., & Kulkarni, A. (1997). Assessment of influence of land use-transportation system on travel behavior. Transportation Research Record, 1607, 105-115. |
| Messenger, T., & Ewing, R. (1996). Transit-oriented development in the Sunbelt. Transportation Reseach Record, 1552, 145-152. |
| Mokdad, A. H., Bowman, B. A., Ford, E. S., Vinicor, F., Marks, J. S., & Koplan, J. P. (2001). The continuing epidemics of obesity and diabetes in the United States. Journal of the American Medical Association, 286(10), 1195-1200. |
| Mokdad, A. H., Serdula, M. K., Dietz, W. H., Bowman, B. A., Marks, J. S., & Koplan, J. P. (1999). The spread of the obesity epidemic in the United States, 1991-1998. Journal of'the American Medical Association, 282(16), 1519-1522. |
| Morrow-Jones, H. A., Irwin, E. G., & Roe, B. (2004). Consumer preference for neotraditional neighborhood characteristics. Housing Policy Debate, 15(1), 171-202. |
| Moudon, A. V., Hess, P. M., Snyder, M. C., & Stanilov, K. (1997). Effects of site design and pedestrian travel in mixed-use, mediumdensity environments. Transportation Research Record, 1578, 48-55. |
| N.C. Center for Health Statistics. (2005). Behavioral Risk Factor Surveillance System. Retrieved January 30, 2005, from http://www.schs .state.nc.us/SCHS/brfss/index.html |
| National Center for Chronic Disease Prevention and Health Promotion. (2004). Behavioral Risk Factor Surveillance System. Retrieved July 28, 2004, from http://www.cdc.gov/brfss/about.htm |
| Northridge, M. E., Sclar, E., & Biswas, P. (2003). Sorting out the connections between the built environment and health: A conceptual framework for navigating pathways and planning healthy cities. Journal of Urban Health, Bo, 556-590. |
| Owen, N., Humpel, N., Leslie, E., Bauman, A., & Sallis, J. F. (2004). Understanding environmental influences on walking: Review and research agenda. American Journal of Preventive Medicine, 27(1), 67-76. |
| Pate, R., Pratt, M., Blair, S., Haskell, W., Macera, C, Bouchard, C., et al. (1995). Physical activity and public health: A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. Journal of the American Medical Association, 273 (5). 402-407. |
| Saelens, B. E., Sallis, J. F., Black, J. B., & Chen, D. (2003). Neighborhood-based differences in physical activity: An environment scale evaluation. American Journal of Public Health, 93(9), 1552-1558. |
| Saelens, B. E., Sallis, J. F., & Frank, L. D. (2003). Environmental correlates of walking and cycling: Findings from the transportation, urban design, and planning literatures. Annals of Behavioral Medicine, 25(1), 80-91. |
| Sallis, J., Kraft, K., & Linton, L. (2002). How the environment shapes physical activity: A transdisciplinary research agenda. American Journal of Preventive Medicine, 22(3), 208. |
| Sallis, J., & Owen, N. (1999). Physical Activity and Behavioral Medicine. Thousand Oaks, CA: Sage. |
| Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Boston: Houghton Mifflin. |
| Song, Y., & Knapp, G. J. (2004). Measuring urban form: Is Portland winning the war on sprawl? Journal of the American Planning Association, 70(2), 210-226. |
| Stokols, D. (1992). Establishing and maintaining healthy environments: Toward a social ecology of health promotion. American Psychologist, 37(1), 6-22. |
| Troped, P. J., Saunders, R. P., Pate, R. R., Reininger, B., & Addy, C. L. (2003). Correlates of recreational and transportation physical activity among adults in a New England community. Preventive Medicine, 37(4), 304-310. |
| Trost, S., Owen, N., Bauman, A., Sallis, J., & Brown, W. (2002). Correlates of adults' participation in physical activity: Review and update. Medicine Science and Spans Exercise, 34(12), 1996-2001. |
| U.S. Department of Health and Human Services, Office of the Surgeon General. (2001). The Surgeon General's call to action to prevent and decrease overweight and obesity. Rockville, MD: Author. |
| U.S. Department of Transportation, Bureau of Transportation Statistics. (2003). Highlights of the 2001 National Household Travel Survey (No. BTS03-05). Washington, DC: Author. |
| [Author Affiliation] |
| Daniel A. Rodríguez is an assistant professor of city and regional planning at the University of North Carolina, Chapel Hill. His research focuses on the connections among land use, transportation, and health. Asad J. Khattak is a professor of city and regional planning at University of North Carolina, Chapel Hill, and director of the Carolina Transportation Program. His research interests include innovations that lead to improved transportation. Kelly R. Evenson is a research associate professor of epidemiology at the University of North Carolina, Chapel Hill, in the Department of Epidemiology. Her research on physical activity epidemiology includes work on understanding the relationships among environment, policies, and physical activity. |
My Research
