[sustran] Re: Accidents in Urban Areas

Paul Barter tkpb at barter.pc.my
Thu Sep 4 20:27:00 JST 1997


Here is an interesting response to Chris Zegras' request which was posted
in the alt-transp list.
Paul.


From: Peter Jacobsen <JACOBSENP at delphi.com>
Date: Thu, 04 Sep 1997 00:30:17 -0400 (EDT)
Subject: alt-transp Re: Accidents in Urban Areas

Christopher Zegras <chris at mailnet.rdc.cl> wrote:

> I am looking for data on traffic accidents in urban areas of both
> developing and industrialized countries. Any leads, references, personal
> papers, data on this topic would be most appreciated.

I strongly encourage you to read about Smeed's Law. Smeed determined in 1949
that the number of deaths per 10,000 registered motor vehicles and numbers
of vehicles per person were related by the formula:

       D      0.0003
     ~~~~  = ~~~~~~~~
       N     (N/P)^2/3

Smeed showed this relationship worked with 20 different countries. John
Adams (1987) then showed this relationship worked in a variety of countries,
throughout time (e.g. Great Britain, 1909 to 1973). Adams also showed it
worked for 62 countries. Adam (1995) graphes the data.

The point is that "accident statistics do not measure safety or danger; as
traffic increases, the death toll is contained, and sometimes reduced, by
behaviour that avoids danger rather than removing it." (Adams, 1995)

Even modern, "safe" vehicles driven in third world countries generate a
death rate that can be predicted by Smeed's law. It isn't safe vehicles nor
safe roads that reduce the death rate -- it's just people adjusting to the
dangers of motor vehicles.

Peter Jacobsen
Pasadena, California

References:

Smeed 1949. Some statistical aspects of road safety reserach. Royal
Statistical Society, Journal (A) CXII (Part I, series 4). 1-24.

Adams 1987. Smeeds Law: some further thoughts. Traffic Engineering and
Control (Feb) 70-73.

Adams 1995. Risk. London, UCL Press



More information about the Sustran-discuss mailing list