Reliability of an Audit Tool for Systematic Assessment of Urban Alleyways

Background:Reliable audit tools are needed to examine the potential of built environment features for physical activity. Methods:An audit tool for alley environments was developed with land use, substrate, and use, condition, and safety items. Two audit teams independently audited 29 Los Angeles alleys, and interteam reliability was calculated with Cohen’s and prevalence-adjusted, bias-adjusted kappa (PABAK) statistics; intraclass correlation coefficients; and percent observed agreement. Results:Forty-two of 47 dichotomous items analyzed for reliability had PABAK values ≥ 0.61 (“substantial agreement”). Sixteen of 23 ordinal and continuous response items analyzed had ICCs ≥ 0.61, and an additional 6 with lower ICC values had observed agreement ≥ 79%. Items concerning the presence or absence of use-related alley features demonstrated the lowest reliability. Conclusions:The instrument has acceptable reliability for most of its items and appears to be a promising tool for use by other researchers and profess...