One vital source of information about what is happening on the highways and byways of Great Britain is offence information.

Two particularly interesting articles in the Interesting Headline Database are:

- Speed offences and average speed cameras

- Drink Drive

The thing to note is that both of these concern records of offences, and that (again) this is very much a byproduct of a system. The number of speed offences depends crucially on the tolerance the police have been using when setting up speed cameras - so the number of offences can vary as the tolerance vary.

Reports of Drink drive offences are also interesting. If in one Christmas Campaign you carefully targetted a location and time of day where you thought there would be a lot of drink drivers, you would end up with the situation where you carried out a modest number of tests, and found a reasonable number were positive. If the following year you carried out blanket wide testing, you would carry out a huge number of tests, and could also find that a reasonable number were positive. But:

1. reasonable number of positive results / modest number of tests = a large percentage
2. reasonable number of positive results / huge number of tests = very small percentage

So, the actual number of drink drivers could be unchanged, but the percentage of tests failed could go down simply because the number of tests goes up.

Task: Have a read of the two articles noted above, and any others of relevance, and think about how the methods employed to assess offences may produce limitations when interpreting findings.