Is Cycling a Safe Mode? Comparing Apples with Apples

Abstract

In many countries, one of the barriers to acceptance and encouragement of cycling by politicians, policymakers, and the public is the perceived danger of the mode. This is not helped by "official" comparisons of different travel modes that typically show cycling as having a far greater injury crash rate (e.g. deaths/injuries per kilometre travelled) than other modes such as driving and walking. For example, in New Zealand, cycling is assessed as having a crash rate per km travelled nine times greater than driving. On this basis, some research and policy reports have warned against people shifting from "lower risk" modes (like driving) to "higher risk" modes (like cycling).

Such assessments may be quite unfair, from the point of view of not "comparing apples with apples". Simple comparisons of overall data ignore a number of factors, including different crash reporting rates, the relative amount of time taken by each mode to travel the same distance, the age distributions and "road experience" of travellers, and the types of road environments or facilities where people are travelling. It also ignores the cumulative "safety in numbers" effect, whereby more people cycling actually helps to reduce the individual risk of each person, and also fails to consider relative health costs other than road crashes, e.g. due to physical inactivity and motor vehicle emissions.

This paper explains how these factors affect the relative safety of different travel modes. Road crash data from 2003-09 and corresponding data from the New Zealand Household Travel Survey was used to compare the relative crash rates for travel modes in New Zealand. This results in quite different (and in many cases much more favourable) conclusions about the relative safety of cycling. For example, per hour travelled, cycling had a similar crash rate for 15-19 year olds and over 75’s when compared to driving a light motor vehicle, even before health considerations were accounted for. There were notable differences in the relative risk of modes when comparing urban vs rural roads and major vs minor roads. In similar-speed environments, a “safety in numbers” effect was also observed for cycling. The results have implications for how crash risk data should be presented and used by public agencies. It also helps to identify the road environments where protection from traffic for people cycling would be most beneficial.