The Safety Camera Debate

A hot subject in the UK at the moment is the use of safety cameras to enforce speed limits.  I find the case for using safety cameras to be overwhelming (despite having been caught twice myself).  The opposing case seems to be a blend of poor logic and poor statistics.

It is easy to forget what the issue really is.  This is not about whether we should have speed limits.  It is about the best way of enforcing them.   There are really only two methods available – cameras or police in cars.  Cameras are far more cost effective. If speed limits are too draconian then surely the answer is to change the speed limit?  It is an extraordinary reaction to a lawy we don’t like to argue that  we should therefore enforce it in a less efficient way.

Once you grasp this then all the research into accident rates with and without safety cameras can be recognised as irrelevant. What we need to know about safety cameras is how cost effective they are in causing motorists to respect speed limits.  If they do this successfully  and the accident rate does not improve significantly then the speed limits are too low and can be raised without endangering people.

As well as answering the wrong question, the anti-camera research into accident rates and speed cameras is frequently very poor. Take this example which the anti-camera faction has been touting recently.  The statistical methods used in the last part of the report which purport to show that speed cameras increase accidents could be used as an undergraduate case study in how not to do significance testing.

The report looks the curve relating accident rate to time and does a Chow test to show that the accident rate was reducing faster before 1990 than it was after 1990 – which is roughy when the first safety camera was introduced.

Error include:

1) Data mining

You could calculate a Chow statistic at any number of break points on the curve and get a “statistically significant result”.  There is actually nothing that special about 1990.

2) Confusing correlation with causation

Even if you to make a convincing case that the different slopes before 1990 and after 1990 were somehow especially important there is no way you can conclude it was caused by the introduction of speed cameras. There are any number of changes happening all the time that affect the road casualty rate – speed limits, car technology, drink driving campaigns, average age of motorists etc. What the report does is like trying to demonstrate that adult literacy courses are useless because the employment rate did not go down when they were introduced.

3) Unjustified exterpolation. The line that purports to show the accident rate if cameras had not been introduced is just silly. If it were true then had we avoided speed cameras we would be getting negative accidents about now. The accident rate is bound to level out whatever technology is introduced – you can’t get lower than zero.


0 Responses to “The Safety Camera Debate”

  1. Leave a Comment

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: