What irks me about all this is the story in the Otago Daily Times, Mr Cliff is quoted as saying: "Our research shows that when police combined high visibility tactics with a reduced speed threshold during Queen's Birthday Weekends in 2010 and 2011, the total number of fatality/injury crashes reduced by 25 per cent, compared with the previous two years. That's an average of 30 people whose lives were saved."
So firstly I'm assuming that this 25% reduction was statistically significant, otherwise it's meaningless (although it seems to be a remarkable coincidence for it to be 25%/one quarter EXACTLY).
Secondly, why are we not comparing 2010, 2011, 2012 and 2013 to the previous four years to get a more robust sample size? Probably for the same reason that Queens Birthday weekend was chosen over Easter, Waitangi, Christmas and New Years, or any other holiday weekend with lots of traffic.
Because their stats wouldn't have worked out any other way… I see this all the time in biological sciences and it's wrong. They are clearly picking and choosing their data to support the bullshit story they want to feed to us. If they were scientist they would be found out and prosecuted for fraud!
How did the weather differ in those weekends? How did the number of vehicles on the road differ in those weekends? There are so many confounding variables it's not even funny and because you can't control for these things you would need data from hundreds of weekends with weather conditions, road conditions, number of vehicles, etc recorded and controlled for to even start to provide evidence that police tactics caused a lower road toll.