Perceived environmental features that could influence
PAP were evaluated by 10 questions about park
facilities, safety, social behavior of other users and
accessibility. The responses originally encompassed 4
categories (“strongly agree”, “agree”, “disagree” and
“strongly disagree”). During analysis, the responses
were grouped into “agree” and “disagree”. These questions
were developed in a specific project that analyzed
environmental and psychosocial correlates of adolescent
physical activity. Focus groups were conducted to
identify specific content to scales that were pilot-test for
test-retest reliability check. Kappa statistics ranged from
0.66 (Poor lighting) to 0.90 (Hard to get to the park).
Sample was described by absolute and relative frequencies
and chi-square tests for proportions. After univariate
analysis, environmental features were included
in a three-level multivariate logistic regression according
to a previously established hierarchical model to test
association with the outcome (PAP practice).