Criminal statistics

Polly claims today, in a piece about the recent local elections, that we should:

Take Hammersmith and Fulham: it boasted the biggest fall in crime

It is a claim she has made before, and it wasn't sourced or qualified then either.

At the time of writing, as was the case before the local elections, this claim is not supported by the Metropolitan Police's Crime Figures website which, as of 9 May 2006, shows that Hammersmith & Fulham had seen a 4.7% decrease in crime on a rolling 12-month basis.

Good. Very good, even. But not as good as Barnet's 8% drop, Sutton's 5.1% drop, Wandsworth's 4.8% drop, or Lambeth's 8.3% drop, and so on.

Of course, it is probably possible to find or construct some measure of crime on which Hammersmith & Fulham has seen the largest fall over a carefully chosen period. However, due to Toynbee's cavalier use of the facts we don't know whether she has carefully chosen the definition and the time period to suit her thesis, or whether she has just made the statistic up.

Sources, please!