Evaluating the Safety In Numbers Effect for Pedestrians at Urban Intersections

Key takeaway:

• The Safety In Numbers phenomenon, which finds that as pedestrian traffic goes up, so does pedestrian safety, has been studied. This paper examines pedestrian and crash data at several intersections in Minneapolis, Minnesota between 2000 and 2013 to see whether the Safety In Numbers effect is observable. The researchers chose to look at the intersection level because intersections are where pedestrian/car interactions are highest.

Results:

• There are existing datasets of crashes, pedestrian counts, and automobile traffic flows, which help inform urban and transport planning projects to improve pedestrian safety. More datasets will become available, especially as technology better facilitates counting and crash detection, which will make it easier to identify problematic areas. The increasing availability and utility of data underscores the need for better standards of practice in data collection, management, and distribution.
• The Safety In Numbers effect occurs at intersections for both pedestrians and cars:
  • Pedestrians were at a lower risk of being hit by a car at intersections with more pedestrian traffic.
  • Cars were at a lower risk of hitting a pedestrian at intersections with more car traffic.
  • Most intersections where pedestrians are of higher risk for crashes are urban arterials with more than one lane of traffic in each direction. The dangerous nature of such arterials is clear to an observer standing at or looking at a photo of such arterials. However, by looking only at data sets, it may be less obvious why these areas are more dangerous.
  • Further work can investigate causality of the Safety In Numbers effect, which is still unknown. i.e., Are pedestrians safer because there are more pedestrians around? Or, are more pedestrians around because it’s safer?

Implications:

• There is no implicit link between more data and better decision-making. More data is useful for improving pedestrian safety insofar as it is legible and makes sense to practitioners and the public. 
• It is important to recognize the need for more data while also recognizing the limits of data. For instance, crash rates and other quantifiable metrics can indicate dangerous intersections, but so can visual and spatial knowledge (i.e., simply being at and observing a particular intersection). Moreover, unlike first-hand, local experience or visual/spatial knowledge, crash data doesn’t measure perceptions and experiences of safety, such as stress levels of crossing a particular intersection, the rate at which vehicles closely pass pedestrians and induce fear or stress, etc.
• A recommended approach is combining crash data with first-person observations, walk audits, intersection analyses, etc., so that data complements qualitative information about pedestrian experiences.

Methods:

• The researchers analyzed and mapped existing datasets on pedestrian traffic counts (to measure pedestrian flow), Average Annual Daily Traffic counts (to measure traffic flow), turning movement counts, and traffic crash records from 2000 to 2013 at 488 intersections in Minneapolis, MN. Through a combination of statistical analysis, geocoding and spatial analysis, they investigated whether and how observable the Safety In Numbers phenomenon was.

Murphy, B., Levinson, D., and Owen, A. (2017). Evaluating the Safety In Numbers Effect for Pedestrians at Urban Intersections. Accident Analysis and Prevention, 106.