BACKGROUND
Crime and safety are well-recognized as potential barriers to outdoor physical activity such as use of parks and trails, especially in low income neighborhoods populated by racial or ethnic minorities where crime rates and the prevalence of obesity and related health conditions may be higher. Previous research (Lindsey et al. 2006) reported models that explain 80% of the variation in daily trail traffic. These models incorporate weather, temporal, neighborhood urban form and socio-demographic variables, and trail features, but do not include objective or subjective measures of crime. This paper explores the relationships between risk of criminal victimization (actual and perceived) and trail use by incorporating crime variables into existing models.

OBJECTIVES
Our primary aim is to explain variation in traffic on multiuse urban greenway trails in relation to crime in trail neighborhoods. Our secondary aims are to (a) extend a working model for forecasting trail traffic; (b) present objective measures of crime problems related to trails; (c) measure the variability of trail traffic in response to different measures of crime (e.g., violent versus nonviolent crime); and (d) quantify perceived safety in proximity to trails.

METHODS
Basic methods for modeling trail traffic are reported in Lindsey et al. 2006. Objective and subjective measures of crime are developed using police administrative records, geographic information systems (GIS), and survey research. Uniform Crime Report data and local law enforcement incident data from 2003 - 2006 were obtained from the City of Indianapolis Police Department. GIS was used to map indices of crime for buffer areas and weighted density surfaces of different sizes (e.g., 0.25 to 2.0 miles) near 23 trail monitors. Multiple regression techniques (negative binomial models and hierarchical linear modeling) are used to extend the trail forecasting model.

Standard survey research methods were used to obtain survey results. The Marion County Obesity Survey (MCOS) was a telephone survey of a random sample of 4,780 adults in Marion County. The ALR-sponsored Indianapolis Trail Use Screening Survey was a telephone survey of 3,197 adults in census tracts within a one-half mile buffer of the trail system that obtained information about trail use and recruited participants for a longer mail survey, the Indianapolis Trail Use and Physical Activity Survey (ITMS; n=465). Both the MCOS and the ITMS included questions about perceptions of crime and trail use and strategies used to avoid victimization. Responses to these questions are used to develop indices of perceived safety or risk.

RESULTS
Crime levels vary considerably among trail monitor neighborhoods in the sample in some unexpected ways. Exploratory models indicate that levels of crime are significantly and inversely correlated with trail use. These results are consistent for various indices of crime (e.g., total crime incidents, numbers of violent crimes) constructed for different spatial areas in neighborhoods around the trails. Compared to previous models estimated for 30 locations, the inclusion of crime measures does not change the significance of the other correlates of trail traffic but the signs on some correlates change direction.

Among respondents to the ITMS, larger percentages indicate trails are “extremely” or “quite safe” than “slightly safe” or “not at all safe”, and very few people reported criminal victimization while using a trail. A substantial minority said “concern about crime” has limited their use of the trails; however, a large majority of respondents report taking “precautions to avoid being a victim of a crime.” Use of trails and perceptions of safety vary by race.

CONCLUSIONS
Traffic levels on urban trails are inversely correlated with levels of crime in neighborhoods surrounding the trails. This finding is robust with respect to various indices of crime constructed at various spatial scales. Although few people report victimization, most take precautions while using trails to avoid crime. Use of trails and other urban pedestrian infrastructure may increase through reduction of crime and better management of fear of crime. Limitations of the current investigation are associated with the limited number of sites in the analyses and inherent limitations of the crime data. Policy-makers in planning, health policy, and criminal justice can use findings to design and assess proposed interventions. For example, health benefits related to increases in physical activity may provide additional justification for crime interventions.

Reference:
Lindsey, G., Y. Han, J. Wilson, and J. Yang. 2006. “Neighborhood Correlates of Urban Trail Traffic.” Journal of Physical Activity and Health, Vol. 3, Suppl. 1, S139-S157. (http://www.activelivingresearch.org/downloads/jpah_10_lindsey.pdf).