Statistical spin of the day

So we get a press release today from the Met, entitled: "Mobile phone crime in Capital falls significantly - Latest from New Scotland Yard". Interesting stuff, right? Perhaps an indicator of some kind of new-found handset equality between the rich and poor, or a pointer to the market being so saturated that phones just aren't worth stealing anymore?

Well, not really. In December 2003 the Met formed the National Mobile Phone Crime Unit (NMPCU), and here are the statistics that, I suppose, should show how well it's worked. Between 1st April 2003 and 31st March 2004 the number of reported "mobile phone offences" was 120,498. The same period in 2004/2005 saw 107,807 offences. And the same period in 2005/2006 saw, er, 108,177. That's a rise, shurely?

But wait! What the Met is really pointing to is the fact that, from April 1st to September 31st 2006, there have been only 52,708 offences, "a five per cent reduction on the same period last year and a 17.5% reduction on the year when the unit was formed".

Now, I'm no statistical genius (in fact I avoided carrying my undergraduate psychology degree to full term simply because I was terrified of having to do a year of stats), but I reckon it's pretty wonky reasoning to: a) use the financial year where the NMPCU was formed two thirds of the way through as the basis for any solid comparison; and b) claim that a 5 percent reduction in cases based on the last six months compared with the same period a year ago is "significant" when the overall year-on-year trend actually shows the statistics rising!

And that's before we even get to the definition of "mobile phone offence". As the NMPCU was set up to tackle "those involved in
mobile phone criminality, focusing on the handlers, re-programmers and exporters of stolen mobile phones", it's hardly the muggings etc that you would expect to inspire the headline "Mobile phone crime in Capital falls significantly".

Or am I wrong?