Methods: 139 males (10-18 years) were enrolled with random digit dial in Philadelphia from 2009-2012. Using GIS-assisted interviews, subjects mapped the path of their activities and reported their perceived safety. Activity paths were overlaid on mapped layers of the built and social environment. Our key predictor was social trust, built as a composite Census tract-level variable from several indicators of social trust measured with a city-wide telephone survey. Mixed effects regression estimated adolescents’ perceived safety based social trust level in areas they traversed, controlling for age, companions, and other confounders.
Results: Perceived safety levels ranged from 1 (low) to 5 (high) and most (78.4%) subjects felt less than entirely safe (<5) at some point. Compared to spending time in areas in the lowest tertile of social trust, participants were 73 percent more likely to feel unsafe (p < 0.01) when in areas with medium social trust and 89 percent more likely to feel unsafe (p < 0.01) when in areas of high social trust.
Implications: The results from our study call into question the assumed protective nature of social trust. Further, these results are salient for community interventions in areas that are both shared public spaces and residential neighborhoods (i.e. public parks in central cities). Understanding the nuanced relationship between social trust and perceptions of safety has important implications for both policies and community interventions addressing crime and violence prevention.