Moving beyond Self-Report: Neighborhood Disorder, Safety and Physical Activity

Stephen Mooney, Columbia University
Michael D. M. Bader, American University
Katherine Bartley, New York City Department of Health and Mental Hygiene
Gina Lovasi, Columbia University
Kathryn Neckerman, Columbia University
Julien O. Teitler, Columbia University
Daniel Sheehan, Columbia University
Andrew Rundle, Columbia University

Lack of neighborhood safety may inhibit physical activity. Stronger associations with activity have been observed using self-reported safety than using independently recorded measures such as crime rates. This may reflect relevance of between-individual differences in perceptions of safety for physical activity, or same-source bias (correlated errors), or an environmental characteristic such as physical disorder undermining both perceived safety and physical activity. We combined self-reported measures of neighborhood safety and physical activity from 509 adult residents of New York City, accelerometer measures of physical activity, and a neighborhood disorder measure developed from systematic observations of Google Street View imagery. Neighborhood disorder was not associated with self-reported or accelerometer-measured physical activity. Perceived lack of safety was associated with 33% greater odds (95% CI: 1.11-1.59) of reporting no physical activity but not with lower levels of objectively measured physical activity. Perceived lack of safety’s association with less reported activity may reflect same-source bias.