Influences of Neighborhood Crime on Mode Choice

Abstract

It is understood that the threats posed by possible criminal activity in a neighborhood can play a major role in the decision to drive, take transit, walk, or ride a bicycle, but little empirical evidence has yet been gathered to support this notion, let alone guide public infrastructure investments, land use planning, or the allocation of police services. This study found that high-crime neighborhoods tend to discourage residents from walking or riding a bicycle. When a high-crime neighborhood is compared with a lower-crime neighborhood, the odds of choosing walking over driving decrease by 17% for work trips and 61% for nonwork trips. However, transit trips appear to respond to neighborhood crime in a way similar to auto trips; that is, high-crime neighborhoods appear to encourage transit use. The odds of choosing transit over the auto increase by 17% for work trips and 164% for nonwork trips. The study found that transit access trips (walking, bicycling, or driving to a transit station) are sensitive to neighborhood crimes in keeping with theoretical expectations: high-crime neighborhoods discourage walking and bicycling and encourage driving to transit stations. Specifically, for transit access to work trips, the odds of choosing to walk or bike to a transit station instead of driving decrease by 48%.

Do characteristics of walkable environments support bicycling? Toward a definition of bicycle-supported development

“…Utilitarian bicycling was positively associated with individuals' perceived number of recreation destinations in the home neighborhood (Beenackers et al 2012) and with bars and drinking establishments , and bicycling for all purposes was positively associated with the number of office/fast food/hospitals and the perceived presence of grocery shops and schools at the home neighborhood level (Vernez-Moudon et al 2005). Bicycle mode choice was positively associated with educational uses at destinations (Winters et al 2010), an origin accessibility to retail and services index for recreation trips (Rajamani et al 2003), and a transit accessibility measure at the household TAZ for commuting trips (Ferrell and Mathur 2012). However, these results do not hold across other studies.…”

“…Some have found a significant but modest positive effect of population density on bicycle commute mode choice (Rybarczyk and Wu 2013) and levels of use (Beenackers et al 2012), while many more have not found any relationship with population density (Frank et al 2008;Schneider 2011;Cervero et al 2009;Rajamani et al 2003;Winters et al 2010;Zhao 2014). Ferrell and Mathur (2012) found a negative relationship between population density in the destination TAZ and bicycle mode choice for work and non-work trips. Employment density has been less studied.…”

“…Measuring density at this scale has become the norm in studies of this kind, but we suspect that this is influenced mainly by pedestrian research, where selection of this scale is deliberately aligned with trip distances and access to destinations. The studies reviewed that did find significant associations between density and bicycling measured the density variable at larger scales of the TAZ (Ferrell and Mathur 2012) and 3-kilometer /1.9-mile buffer (Rybarczyk and Wu 2013), which are more commensurate with distances traveled by bike.…”

“…• A pedestrian/bike factor at trip destinations based on intersection variables (Cervero and Duncan 2003) • Intersection density at trip origins or households (Frank et al 2008;Ferrell and Mathur 2012; Zhao 2014) • Intersection density along routes (Winters et al 2010) • Connectivity measured through network intensity attributes of urban morphology 1 at the home neighborhood for commuting (Rybarczyk and Wu 2013) Two more bicycling studies (e.g., whether a person bicycles at least once a week) found that utilitarian cycling was positively related to street density (Cervero et al 2009) and recreational cycling was also positively associated with connectivity around the home (Beenackers et al 2012).…”

“…This is also true of trip distance or travel time variables. In three US studies, the magnitudes of the trip distance coefficient for walk mode choice were more than three times those of the corresponding coefficient for bicycling (Appleyard 2012;Cervero and Duncan 2003;Ferrell and Mathur 2012). …”