There was a bit on the
BBC South Today about this. Oxfordshire CC switched off all their speed cameras last week but 'accidentally' left a few on. One camera recorded 62 speeders in a 5 day period, an 88% increase for the first week after the announcement. Another recorded 110, 18% more than the average. The council will not be prosecuting any of the drivers.
Just to blat some simple stats at this one.....
62 as an 88% increase implies that the mean number of speeders over this period is around 33. Using the poisson distribution (may not be completely accurate, but good for counting event type experiments) , there's about a 1 in 10,000 chance of this happening accidentally (99.99% upper confidence interval - even tails - on a Poisson with mu=33 is 61.5).
In the interests of fairness, the same analysis for the 18% increase yields about a 1 in 10 chance of the variation being purely statistical.
I'm not going to go all Paul Smith (RIP) about these fag-packet sums, but I can't imagine the council has any motivation to perform a long term study at ex-camera sites to establish some proper measurements.
Of course, we could just appeal to people's nice side by putting up a sign that says 'please' and possibly lights up, but I suspect that these were only effective because drivers were unsure whether there was a camera attached as well.