Fear of Crime

The ONS has suggested there’s a big gap between fear of crime — using robbery as an example — and the likelihood of being a victim. Is it right?

The article says: “0.3% of adults were victims of robbery in the year ending March 2016, but 9% of those surveyed were very worried they would experience it in the forthcoming year — 30 times higher than the rate of victimisation.”

Is anything wrong with that? Ask yourself this question: if 1000 people knew that three of them would be robbed sometime in the next year, perhaps as they walked home from the station on a winter evening, perhaps outside a bar, how many should be seriously worried?

There are several possible answers.

You might say ‘all of them’, as everyone can wonder: ‘will it be me?’

Or you might say that everyone should be a bit worried, since being robbed is bad but, because it’s unlikely, that mitigates the worry somewhat.

Or you might say there are no rules about who should or shouldn’t be serioulsy worried, since it depends how horrible is the prospect to each of us of being robbed, and how content we are with average odds of more than 300–1

What I’m pretty sure you can’t say is that only three should be seriously worried because that is proportionate to the risk. But isn’t that what the ONS has implied? There was even a graphic to suggest how silly we are.

You could take this to imply that that the public is not just wrong but irrational. But if only three people in a thousand were seriously worried, the chance that the three who were worried would be precisely the three who were robbed is vanishingly low. The victims would most likely not be them.

There is simply no way of tying the degree of risk in this case to who or how many should be deeply worried. They are different things, apples and pears. Because 126 people have been reported killed by terrorism between 2000 and 2017 in the UK, should just 126 should fear it in the next 16.75 years? Which 126 exactly? When it comes to being robbed, there is only one deeply implausible scenario in which three would be the right answer: if the three who worried knew they were going to be the three victims, and they were.

Lacking clairvoyance, if we want to say that fear of crime is disproportionate, we have to be careful what we mean. If we mean that more people are worried than will be victims, well, that’s not surprising. It’s why people have insurance. Worry is partly an expression of uncertainty, not only frequency. It’s the degree of worry that might be disproportionate, such as extreme anxiety over an extremely rare event that has limited impact.

The report was written to launch the ONS crime calculator, which received good coverage — like this from the BBC — and is a useful tool for checking the likelihood of crime for different demographic groups.

The crime calculator is handy. And it could be used to make a more careful point: how much crime do people think there is versus how much crime there really is. The BBC’s coverage of the crime calculator wasn’t quite free of the worried/at-risk problem, but it made some good points, such as how variable the risk is depending on our characteristics (the calculator can be personalized), and that the young are less likely to be worried than the old, even though more likely to be victims. The latter, though, illustrates our point: who is right, young or old, both or neither? What would be the correct number of each?

So, it may be that better knowledge of the likelihood of crime will change some people’s worries if they hear (and believe) that the likelihood is much lower or higher than they thought. For that reason, the crime calculator might help. It is, in any case, a good thing to have the data in a neat, accessible form, especially personalized.

But we simply can’t align ‘per cent worried’ with ‘per cent harmed’.