People always claim statistics are inaccurate when the statistics contradict their subjective experience or sincerely held belief.

One proxy for crime statistics is homicide. Any other stat can be juked, but homicide cannot because bodies stink. To someone who was insisting to me that the UK crime rate has increased over time, I pointed out that it was possible, but unlikely. Homicide is at a multi decade low in absolute numbers, and much lower in per capita numbers. That’s likely to be correlated with violent crime.

So in the case of the UK both homicide and violent crime have followed a gentle downwards trend line, and we can trust the first line completely. Therefore, the second is likely to be true.

Did it convince the guy I explained this to? No it did not. “You don’t live in a rough area like I do”. Ok then.