Brownson, R., Hoehner, C., Brennan, L. K., Cook, R., Elliot, M., & McMullen, K. (2004). Reliability of Two Instruments for Auditing the Environment for Physical Activity. Journal of Physical Activity and Health, 1, 191-208.
PURPOSE: To understand the relationships between street-scale environments and rates of physical activity, it is crucial to develop reliable methods of measurement. Community audits are commonly used to test the walkability and bikability of environments, yet few have been tested for reliability. METHODS: Audit tools were collected from the peer-reviewed literature, the Internet, and experts from a variety of backgrounds. Two versions of an audit instrument were created: an “analytic” (with Likert-scale and ordinal-response choices) and a “checklist” (with dichotomous response choices) audit tool. Audits were conducted in St Louis, MO for 147 street segments, representing both higher and lower income neighborhoods. The same segments were re-audited to assess interrater reliability. RESULTS: Characteristics of the physical environment varied considerably across lower and higher income segments. For example, in the checklist audit, physical disorder was present for 67 segments in lower income segments, compared with 0 segments in higher income segments. Among 8 questions from each audit tool designed to broadly capture environmental attributes, most had moderate to poor agreement. Most of the transportation and land-use items demonstrated high (substantial or perfect) agreement, and the aesthetics and social environment items showed reliability in the moderate to fair range. CONCLUSIONS: A community audit tool can be relatively easy and quick to administer and, for many domains, is reliable. Our audit tools appear particularly well suited for capturing elements in the transportation and land-use environments.