Background
This study compares walking behavior (measured by walking diary), perceptions of the environment as they relate to physical activity and walkability, and moderate intensity and total physical activity (measured by accelerometer), with the physical and socio-structural environmental features of residential environments. Psychosocial variables well correlated with physical activity will also be assessed.
Specific Aims
1. To compare self-reported walking behavior and objectively measured physical activity with the physical and socio-structural environmental features of residential environments. It is hypothesized that individuals from high density, high street connected neighborhoods engage in more walking behavior and thus greater physical activity than individuals from low density, low street connected neighborhoods.
2. To compare self-reported walking behavior and objectively measured physical activity with perceptions of the environment as they relate to physical activity and walkability. It is hypothesized that those who perceive the environment to be more walkable will be those who walk more. Further, we will be able to assess correlations of environmental perceptions with more objective measures of the physical, socio-structural environmental features in neighborhoods.
3. To assess the amount of variability in walking behavior and physical activity that can be explained by physical and socio-structural environmental features, including environmental perceptions and aggregated psychosocial variables such as self-efficacy, social support, and barriers to exercise. It is hypothesized that physical and socio-cultural features of the environment, as well as environmental perceptions of the same, will explain variance in walking behavior and physical activity. Individual-level psychosocial and socioeconomic variables are expected to mediate these relationships.
Methods
Our unit of analysis is residential areas selected from the environmentally diverse northern sector of the Minneapolis-St. Paul metropolitan area, stretching from the urban core to the urban edge, for which especially rich GIS data are available. We will compare walkability in 36 0.5*0.5 mile core study areas selected for variation along two dimensions: (A) street pattern and (B) residential density. Potential core neighborhoods will be ranked high, medium, and low on both dimensions. To maximize variability, we will focus on a subset of the most extreme areas, ranking high or low on each of these two dimensions. Consistent with established psychometric techniques, we will decompose the variability both within and between the areas, with both fixed and mixed effect models, and employ novel but established propensity score methods to better control confounding and estimate "causal" effects of context on endpoint. The dimensions of mixed use and pedestrian infrastructure, and more standard socio-economic factors, will be measured for each area but not used as the sampling frame.
Twenty participants will be recruited from core area. We will repeat GIS-based measures for a 1.0*1.0 mile "context" area, surrounding each core area.
All social and walking behavior measures will be made at the individual level in each residential area. Specific aims will be assessed using established statistical procedures, including ecological correlation and regression, specifically fixed and mixed effect models. Results will form guidance for urban development and a statewide coalition for physical activity advocacy (www.beactiveminnesota.org).
