We hypothesized that for our sample of minority women, a stronger concordance of directly and indirectly measured built environment attributes would be significantly associated with PA adoption. Objectively measured PA did not significantly increase, but self-reported PA did significantly increase from T1 to T2. PA changes over time did not vary by ethnicity or any concordance measure.
No earlier study has measured the association between built environment attribute concordance and PA changes over time, but PA has been reported to be a significant correlate of built environment attribute concordance . In particular, one study found lower concordance among women with lower income, PA and self-efficacy for PA . Also, other findings suggest that indirectly measured neighborhood data are more closely linked to self-reported PA than directly measured neighborhood data [27, 42]. Unlike studies measuring direct and indirect built environment attribute concordance, our sample consisted solely of minority women. The relationships between PA and attribute concordance might differ for our population, as earlier findings suggest that the degree of built environment non-concordance can vary among certain population subgroups . Also, our samples were of high SES, particularly for income and education; we also assessed a wider variety of neighborhood types than previous studies [20, 42], increasing the generalizability of our findings.
Although not all of our participants exhibited increased PA over time or PA adoption, this study initiates an evidence base where no similar data exist. PA adoption is an essential component to a healthy lifestyle [3, 43], yet no known study has measured the associations of PA changes over time with built environment concordance values. Further, this study investigated these relationships among minority women. Although African American and Hispanic or Latina women continue to be disproportionately physically inactive compared to white women [2, 3], they continue to be understudied in the built environment literature .
Other strengths of this study include the use of a self-reported PA questionnaire and accelerometry to measure PA changes over time, providing a comprehensive assessment of PA. Although similar studies have been cross-sectional in nature [20, 27, 44], our study measured PA longitudinally. We also used measured BMI and body fat percentage, rather than self-report, helping to reduce bias and measurement error.
Our study is not without limitations. Due to adherence, cost and logistic reasons, the number of participants who wore accelerometers was lower than those who completed the self-reported PA questionnaire at T1 and T2. Resources are needed for future studies to recruit and assess an equal number of participants for multiple PA measures to provide a more comprehensive PA assessment. McCormack and colleagues found that residents' perceived behavior control cognitions were mediators in the relationship between the built environment and PA , and future work is needed to include additional individual-level variables that might help explain the variability of attribute perception(s) and PA changes among these populations.
This study investigated built environment measurement concordance and PA changes over time among minority women. Inaccurate perceptions of built environment attributes were not associated with PA level change. Future PA interventions and supportive communities could promote built environment attributes (e.g., park amenities, clean baseball fields, long walking trails) in an attempt to increase PA. Policies could attempt to increase facility and street signage in an effort to promote PA, particularly among ethnically diverse neighborhoods.