Chris Grayling was let out of the dog house this morning for a rare media appearance on the Today programme . Challenged on crime statistics, he at least conceded that "overall crime has fallen".
But just as it seemed that we might be making progress, the shadow home secretary repeated his false claim that violent crime has risen dramatically under Labour. In fact, the British Crime Survey (BCS), still regarded by statisticians as the most reliable long-term measure of crime, shows that violent crime has fallen by 41 per cent since 1997.
Conveniently, Grayling refuses to accept the BCS and prefers to use police-recorded statistics, based on individual reports by victims of crime. But as the Violence in England and Wales 2009: an Accident and Emergency Perspective report  shows, even on this measure, violent crime is falling.
As the graph below shows, the BCS, police recorded crime and accident and emergency figures all point to one conclusion: violent crime is falling.
But the empirically challenged Grayling then decided that it wasn't all about statistics after all. In a remarkable claim for a prospective home secretary, he said:
I don't think it's just about figures, I think it's about what people see in their communities. I mean, i do think Britain is a more violent place than it was a decade ago, it's the country where only a few weeks ago an elderly couple died after someone set light to their mobility scooter.
The assertion was clear. Grayling is able to set aside statistical evidence on the basis that his own personal experience, plus one or two anecdotes he's picked up from the tabloids, proves that Britain is a more violent place.
Never mind that the most reliable figures show nothing of the sort, Grayling just knows that Britain is more violent now than it was in 2000.
Should he ever make it to the Home Office (and the odds are against it ), one expects that Grayling, keen to prove that a Tory government has cut crime, won't adopt such a cavalier attitude to the facts.
UPDATE: Over at the Spectator's Coffee House blog, David Blackburn takes me to task  for my claim that violent crime has fallen.
First, he argues that changes in recording practice mean that figures from the most recent British Crime Survey are incomparable with those from 1997. But, as I point out in the comment thread, the change only applied to police recorded statistics (favoured by Grayling), not to the British Crime Survey, which has measured crime in the same way since 1981.
Second, he points to a recent document from the House of Commons library which stripped out 24 per cent of the increase in violent crime to account for the new recording methods, allowing the Tories to claim that violent crime has risen by 44 per cent since 1998.
But what the party failed to mention is that the 24 per cent figure accounts for only one year of the changes, even though the violent crime figures were artificially inflated for at least two to three years.
It's still Grayling who has the explaining to do here.