Presentation at the 2007 Active Living Research Annual Conference
The built environment has been recognized for its potential to encourage children’s walking to/from school, and thereby increase children’s physical activity. However, very few studies have explored the relationships among different aspects of such environmental support. Minority children in the U.S. are at high risk of physical inactivity, obesity, and traffic injury. Yet the disparity issue has not been studied thoroughly in terms of environmental support for children’s walking. Those low-income, minority, urban neighborhoods have been understudied as compared with middle-income, White, suburban neighborhoods.
The aims are: 1) to compare elementary schools with different ethnicity distributions in terms of the environmental support for walking to/from school; 2) to explore the relationships and possible conflicts among various aspects of walkability and safety.
This is a cross-sectional study in the Austin Independent School District of Texas, which features a high percentage of Hispanic students (54.7%) and a wide range of physical environmental features. Geographically, Hispanic children cluster in the eastern area (P < 0.01 for Moran’s I), while the western area has a majority of non-Hispanic White children.
Variables and constructs of walkability and safety are identified from previous literature linking the built environment with children’s walking to/from school or the general population’s walking. The macro-level walkability (i.e. urban forms) is measured by distance to school, pedestrian facilities, residential density, street connectivity, and land use mix; the micro-level walkability (i.e. urban design and architectural qualities) is assessed through functionality, aesthetics, amenities, maintenance, and perceived safety; traffic safety is measured by traffic volume and speed; crime safety is measured by the yearly crime rate.
A Geographic Information System (GIS) is used to measure the macro-level walkability, traffic safety, and crime safety for each school’s attendance area (n = 73). Data sources include the City of Austin’s GIS datasets, U.S. Census 2000 data, Texas Capital Area Metropolitan Planning Organization, Texas Department of Transportation, and Austin Police Department. Field audits are conducted to assess the micro-level walkability for 200-meter street segments sampled from attendance areas (n = 39). The inter-rater reliability is tested by Intraclass Correlation Coefficients.
In the GIS audit, except for traffic volume and speed, all constructs show significant cluster effects (P < 0.01) and certain extent of spatial trends relevant to ethnicity distributions.
An Analysis of Variance is run between the percentage of Hispanic students (categorized according to quartiles) and each environmental construct. As hypothesized, the percentage of Hispanic students is positively associated with constructs of macro-level walkability (except street density), yet negatively associated with constructs of micro-level walkability (except functionality) and crime safety (P < 0.05). There is no significant association for traffic safety.
The estimated mean differences are calculated for the 4th group of schools (≥ 82.1% Hispanic students) and the 1st group (< 37.6% Hispanic students) (P < 0.05). For macro-level walkability, the group with the most Hispanic students has 34.8% more students living within 1/2 mile of school; has 15% more sidewalks; has 1 more traffic signal per 10 miles of streets; has about 427 more residents and 8 more street intersections per 100 acres; and has a 0.165 higher land use mix index on the 0 to 1 range. In contrast, for micro-level walkability, the aesthetics, amenities, maintenance, and perceived safety are all rated at least 1 point lower on a five-point scale. What’s more important, these schools have 162 more criminal offenses per 1,000 persons per year in their attendance areas (mean = 239).
A factor analysis with varimax rotation is conducted for the standardized scores of significant constructs, using 0.6 as the threshold value of factor loadings. Results extracted 3 components, with 4 micro-level walkability constructs in component 1 (all factor loadings > 0.9); distance to school, sidewalk completeness, and population density in component 2; and traffic signal density, street intersection density, land use mix, and crime rate in component 3.
Schools with more Hispanic children feature attendance areas with higher macro-level walkability, yet much lower micro-level walkability and crime safety. Threats from crime tend to grow with the increase of land use mix and street connectivity, which are usually considered to encourage adults’ walking. Therefore, targeted research is needed for the impact of these intertwining features on children’s walking to/from school, especially for low-income, Hispanic neighborhoods. In addition, the micro-level walkability should be studied as an important and holistic aspect parallel with the macro-level walkability. In summary, the promotion of walking to/from school has to recognize possible conflicts among different aspects and scales of the built environment.