Thursday, July 17, 2008

Why trust your own experience when the headlines can tell you what to think?

Crime is down. Again. Violent crime is down, too. But what about public perceptions of crime?

According to the survey [PDF], 65% of people think crime has increased nationwide over the last two years, but only 39% of people think crime is up in their local area.

The difference is even more striking when you distinguish between those who think crime has gone up a little and those who think it’s up a lot. The figures for crime going up a little are pretty close: 26% think this is true locally, 31% nationally. But just 13% of people think local crime is up a lot, while 35% think crime is nationally up a lot. In areas that they don’t have first-hand knowledge of, people are far likelier to think things are going to hell in a handcart.

The conclusion from this is that people’s perceptions of crime nationally – which are influenced far more by the media than by personal experience as compared with perceptions of local crime – are absurdly overinflated. Society isn’t ‘broken’, but the media are. Politicians who feed the frenzy and then base their policies on the resultant ill-informed public fears are doing us a disservice.


M said...

I don’t think you put much, if any, thought into establishing the validity of the crime statistics. If you did you’d realise that one is crimes reported to police, which obviously omits those not reported e.g. those crimes where the victim is to scared to go to the police or believes the police cannot help them. The other is the BCS, which is supposed to be an overall sweep of people’s experiences of crime, unfortunately this caps the experiences of repeat victims e.g. domestic violence victims, people who are terrorised etc, to stop them inflating the calculations, of course I’m sure it’s just coincidental that the crimes which are capped are the sorts of crimes the victims are too scared to report to police.

Tom Freeman said...

I think I probably put enough thought into it. The report I linked to finds that both overall crime and violent crime as recorded by the police are down; it also finds that both overall crime and violent crime as measured by the BCS are down.

Both measures have their weaknesses but they had the same weaknesses last year, when higher crime levels were observed.

The perceptions of crime nationally and locally are a different set of figures entirely. I assume that people have more accurate knowledge of their local areas' crime levels than of national crime, so given the disparity in perceptions I conclude that perceptions of national crime rises are far likelier to be wrong.

John B said...

@ MJW and others of your ilk - "just 13% of people think local crime is up a lot, while 35% think crime is nationally up a lot" is consistent with Tom's narrative that crime is stable but the media is crazy. In your parallel world where crime is really rising but people are too scared to report it, what's the explanation for that data?

M said...

john b, we don't know what is happening to crime, because we've got no truly accurate measure. Local vs national perceptions is interesting, but if the respondents use "national" simply as a substitute for "somewhere else other than where I live" then it's easily explanable.

There is no paralell world scenario, you're just being silly, we simply don't know if actual crime is rising or not, the government can't actually tell us, and probably has no interest in telling us, all it is doing is saying "look here's some dodgy statistics to compare against some dodgy statistics from last year, if you pretent that the level of dodginess has stayed the same then it might tell you the story we want you to hear".

The problem with comparing flawed (and in the case of BCS, manipulated) data sets year on year is that we simply don't know how the change relates to respective impact of the flaws and manipulation in any given previous year. We can't make the asumption that the fiddling of the data carries the same weight year on year, because it's an arbitrary cap not a weighted adjustment intended to keep the impact of the fiddle at a conistent level.

The press may mislead us, but so do the government.