Observed, GIS, and self-reported environmental features and adolescent physical activity.

PURPOSE: Examine associations among observed, self-reported, and Geographical Information Systems (GIS) environmental features and physical activity among adolescent males.
DESIGN: Cross-sectional study.
SETTING: Boy Scout troops and neighborhoods in Houston, Texas.
SUBJECTS: Two hundred and ten 10- to 14-year-old Boy Scouts. MEASURES: Accelerometry to obtain minutes of sedentary, light, and moderate to vigorous activity per day. GIS sources were used to identify the numbers of parks, gymnasiums, trails, bus stops, grocery stores, and restaurants within a 1-mile radius of participant residences as well as residential density, connectivity, and crime. Participants provided a self-report of their environment. ANALYSIS: Principal component analysis was used to reduce the number of GIS and self-reported items. Four factors were previously obtained from direct observations of the neighborhoods. Correlations were conducted among factors and physical activity. Regression models were run in which minutes of sedentary behavior, light, or moderate to vigorous physical activity were the dependent variables and environmental factors were the independent variables. Nonsignificant variables were removed in a backward deletion process.
RESULTS: Three GIS factors, Parks, Crime, and Gyms, were obtained as were two self-reported factors: difficulty and access and safety. Factor scores were interrelated and associated with the four observed factors. Only observed sidewalk characteristics were correlated with physical activity and were retained in the regression models.
CONCLUSION: Environmental factors were interrelated. Only sidewalk characteristics were associated with sedentary behavior and light intensity physical activity.