this from www.safespeed.org.uk, dated June 2004 but interesting nevertheless, in view of today's news of Speed Cameras being vandalised in NSW Speed camera report is false NEWS: For immediate release The recent official report on the benefits of UK speed cameras is totally unjustified in its headline conclusions, says Safe Speed. The largest problem arises from the naturally random distribution of road accidents, and the rule that requires cameras to be placed where there have been a high level of accidents over a fairly short period. When a camera is placed where there has been a temporary upward blip in accident figures, we should not be surprised that the blip does not recur. Such placement appear to be extremely commonplace and the effect results in an illusion of benefit. Even genuine accident black spots are most likely to receive a camera after an especially bad spell. Safe Speed has written to Professor Ben Heydecker, the report's apparent lead author, demanding his resignation and the withdrawal of the report. Paul Smith, founder of the Safe Speed road safety campaign, explains: "This error is well known to statisticians, road safety professionals and the Department for Transport (DfT) as "regression to the mean". In fact the DfT even publish warnings and correction methods in their "Road Safety Good Practice Guide". It is shameful that this error has been allowed to remain, uncorrected and unannounced, in the new official report." Paul continues: "Last year's report contained the same error, and Professor Heydecker admitted as much on the BBC Radio 4 programme "more or less". It is all the more shameful then, that the error has been repeated in the new report." Make no mistake. This is a huge error. In a well known study of 2,500 crossroads in Sweden, it was noted that those junctions left untreated showed an improvement in accidents of more than 50%. If a similar error exists at UK speed camera sites, and it might well, then the truth could be that speed cameras caused a 10% increase in killed and seriously injured accidents at speed camera sites. "The problem with regression to the mean", explains Paul, "is that accidents may not be reducing at all. They simply take place somewhere else. That is exactly what seems to be happening especially to our fatal road accidents." This site has a HUGE resource of safety-related stuff, well worth a long look.