For healthcare professionals only

Hazardous Journey

Car colour and risk of car crash injury: population based case control study

BMJ 2003; 327 doi: (Published 18 December 2003) Cite this as: BMJ 2003;327:1455

Statisticaly worthless

It appears that someone doesn't understand the concept of
percentages. Perhaps, a wander through a fifth grade maths book, is in

The answer to why silver cars seem to crash less often than cars of
another color is easily answered, by driving through any parking lot and
averaging the number of silver cars, against the other colors of cars. (a
large parking lot at a shopping mall makes for a good cross-section of the
driving public)

You will likely find, as I have, that there are more silver cars out
on the roads, than any other color. Because of this, the number of crashes
caused by cars of that color is statistically less. That is, not to say
that fewer crashes happen - just that those crashes are weighed against a
larger whole.

With less-popular colors (brown) the number of crashes becomes
statistically higher, simply because there are fewer of that color of car.

For example, four crashes seems more statistically significant, when
you only have six of a color of car on the roads (2/3) than when you have
twenty (1/10).

I'd like to see the actual totals of the accident numbers reported.
I'd be willing to bet that the number of crashes per color of car is about
the same. The total number of each color of car in their sampling should
be supplied as well.

Without these extra bits of information, this study is essentially

Competing interests:
None declared

Competing interests: No competing interests

12 January 2004
chris t sutor