Safety in Numbers

Comments on this entry are closed.

• 

  Geof Gee Link November 18, 2010, 5:19 pm

  While I believe that the highlighted statement is true — that proportionately more cyclists on the road leads to reduced collision rates — there are also reasons to be skeptical that the relationship in the graph captures the true effect and likely overstates it. This graph of two ratios composed of three variables is subject to a natural bias as described by the working paper below.

  http://tinyurl.com/2bfy95z

  Occasionally, Jacobsen’s 2003 paper comes up and I have written a short post on that as well.

  http://tinyurl.com/238kae4

  Just to be clear, my point is not to rain on everyone’s parade but instead emphasize that we need to think about what these studies mean and if the empirical relationship is true, think carefully about why it is true.

  John S Allen wrote something on that.

  http://tinyurl.com/327omcn

  As you can see, the underlying cause(s) is important. Particularly since — simpy based on anecdotal conversations — Safety in Numbers seems to be used as an excuse to accept cycling oriented engineering that clashes with our common sense, traffic science, and experience on what is safe.

• 

  Allen Muchnick Link November 30, 2010, 11:20 pm

  Geof is correct; correlation is not causality, and “safety in numbers” (aka SIN) is an often-abused slogan. While safety and trip incidence are generally well correlated for all travel modes (including auto), at least some of this correlation results from “numbers in safety”; e.g,, the safer that bicycling is (or seems to be), the more that people will actually ride bikes.

  Nations and communities that actively encourage bicycling have diverse and active programs (encompassing all 6 “E”s) intended to make bicycling both safer and more popular. The graph above is hardly remarkable, whereas the lack of a clear correlation would suggest that strong government bicycling programs have a net negative effect.