No, no, don't worry, I'm not about to wade into which set of crime figures are correct, the survey or what the police write down. Nor even about which political party is tripping over its own feet least in playing football with the figures. No, rather, I just want to address a particular point that's often made.

While the definitions of some crimes may have changed over time, or be either way crimes, murder is murder so that's the gold standard to look at. Changes in the murder rate will therefore be the best guide to whether crime is increasing or not.

This isn't, I sorry to have to say, entirely true. Yes, we can indeed measure murder rather better than we can all other crimes. But we're still missing something. That is that medical treatment of trauma victims has got a great deal better over the decades: people attacked who would have died in earlier times (ie, would have been murdered) now survive (and are thus not murdered). So by counting only the number of people successfully murdered we're confusing two entirely different things. The number of people attacked so that they might be killed and the number of those who survive or succumb to such attacks. Yes, this is just a newspaper report, but the estimates of how important this is are large:

Improvements in emergency care over the last 40 years have helped to lower the death rate among assault victims by nearly 70 percent, a new study says.

Those figures are for 1960 to the turn of the century. Over that time (page 9 here) murders have gone from 300 ish a year to 600 ish a year (one year's figures are no good for one event, a bombing, or Harold Shipman, can change the figures hugely). Population has also changed of course, from 48 million or so to what, 65 million today?

Now quite how you want to crunch all of those figures together is up to you but the number of murders has doubled while population has risen by only 35%...and we would expect, as a result of better medical care, the number of murders (assuming assaults of equal severity taking place) to have fallen substantially.

All of which leads to two points. Murder is, by definition, successfully killing someone and if the rate of success changes then we cannot use the simple number of murders as our standard by which to measure crime rates. And when we adjust by one way of looking at that success rate, the medical care which prevents such success, then it really does look at if Britain has become a much more violent place over the decades. For murder should have fallen and it's risen.