Factcheck: Can we trust anti-social behaviour statistics?

Can we trust anti-social behaviour statistics?

23 September, 2010 - 17:03 -- Full Fact team

Today's HMIC report on anti-social behaviour has been seized on by the media as showing there is an incident of anti-social behaviour once every two seconds. It took us only a little longer than this to start seeing problems with the figures...

Since their introduction in 1998, Anti-Social Behaviour Orders (ASBOs) have regularly attracted derision from various quarters of the media and public opinion.

So when Sir Denis O'Connor, head of Her Majesty’s Inspectorate of Constabulary (HMIC), delivered a damning judgement on the ASBO, the media leapt upon the findings from the report, plundering it for quotes and statistics.

Articles claimed that the report painted the sorry picture of a country where “yobs rule the streets”, and police had “given up”.

For support, they drew on claims about the enormous number of anti-social acts committed in the UK. But were their figures correct?

The Claim

The HMIC report states: “It is estimated that the public only report just over a quarter of incidents of ASB to the police – about 28 per cent. Even this low reporting rate led to around 3.5 million calls to police in 2009-10.”

Several media outlets have combined these two sentences to arrive at a total figure for anti-social behaviour.

The Daily Telegraph, Daily Mail, The Guardian, The Daily Mirror and The Sun, to name but a few, inferred from this sentence that there were 14 million incidents of anti-social behaviour each year, or approximately one every two seconds.

But do the papers’ sums add up? Are the figures on which they are based even sound?

Full Fact’s investigation suggests the answers to both these questions may be ‘no’.

Analysis

The HMIC report does not contain an estimate of the total number of reported and unreported anti-social acts in the UK. But it is not hard to see how reporters arrived at their figures using figures used in the report.

If the 3.5 million incidents reported constituted “around 28 per cent” of the total, this total could be reasonably be estimated at around 12.5 million – close to the figures reported in some papers.

The 3.5 million figure is given in further detail at the back of the report. But how can we possibly have figures for the number of incidents that are not reported?

Full Fact followed the footnotes to the 2007/08 British Crime Survey (BCS), where we found a document on perceptions of anti-social behaviour. But the 28 per cent figure was nowhere to be seen.

There was a survey asking witnesses of specific kinds of anti-social behaviour and whether or not they had reported it.

But when we used the BCS survey to calculate a percentage, it came out as 25 rather than 28 per cent.

This mystery was solved when we discovered that the HMIC report had made basic mathematical errors when reviewing the data.

Even more significantly, only a minority of the complaints that had been made were to the police – the survey also measured responses such as talking to parents, landlords or the perpetrators themselves.

On this evidence, the actual percentage would be radically lower and the estimate for anti-social incidents far higher than that reported in the news.

The problems don't stop there.

There is a clear problem in applying a percentage derived from a survey of people’s reported experiences of some types of anti-social behaviour to a survey based on actual reported anti-social behaviour.

As well as the erroneous implication that the BCS data measures calls to the police, HMIC also suggests that it is a measure of all anti-social behaviour.

It is not. The categories of anti-social behaviour mentioned in the survey do not correspond to police categories or definitions, ignoring such common complaints as abandoned vehicles and “malicious communications”.

Even if the figure of 28 per cent were accurate – which it is not – the report was not intended as an authority on how frequently anti-social behaviour goes unreported, and it is inappropriate to quote it as such.

Some papers used the 28 per cent figure to calculate total incidents, others rounded this to a quarter, producing results of between 12.5 and 14 million. But if this were true, would it really mean an incident every two seconds?

There are 31.5 million seconds in a year, so for an average of an incident every two seconds the total number of incidents would need to be just under 16 million. So not quite.

There are questions not only about whether 28 per cent is the actual proportion of anti-social behaviour incidents that do get reported, but even more questions about how this figure has been applied in the press.

Conclusion

There are serious questions about the way the statistics have been used by both HMIC and the mainstream media.

With warnings about the implications of spending cuts for police power to curb anti-social behaviour, and a new Government keen to make progress on the issue, it is vital that the upcoming debate is served by accurate information. The high level of public interest only intensifies this need.

This analysis has shown that the assertion of 28 per cent as the proportion of total anti-social behaviour incidents that actually get reported is based on a misunderstanding of the statistics, and. for all anyone knows, wildly wrong.

Furthermore while HMIC do not connect this estimate to the 3.5 million reported instances of anti-social behaviour, by putting the two sentences next to each other, it is hardly surprising that the media used the two figures together to get an estimate for total anti social behaviour.

Nevertheless the press went beyond even what the flawed report justifies. There appears to have been some generous rounding up to claim that the figure is 14 million, and even more to say this constitutes an incident every two seconds.

But the real problem is that there just aren’t the statistics available to back up these kinds of claims on anti-social behaviour.

The most recent edition of ‘Crime in England and Wales’, the comprehensive official figures on crime contains the following warning on the use of anti-social behaviour statistics:

“It should also be noted that much ASB be may reported to the local authority (e.g. noisy neighbours) and may therefore not come to the attention of the police. There may also be multiple reporting of the same incident to different authorities (e.g. both police and local council). The BCS has something to offer in terms of public perceptions of ASB but it is certainly the case that our knowledge of the nature and impact of these incidents is not as well understood as for crime.”

Full Fact is now pressing HMIC to correct their report and the press to take the opportunity to make clear to their readers that the figures used in their reports were not accurate.

Edgar Gerrard Hughes & Patrick Casey

Liked this? Read these:

Can't get enough of Full Fact? Sign up to our monthly newsletter

Follow @FullFact

Comment is free but facts are expensive!

Full Fact believes in the possibility of accurate and informed debate. Our factchecks look at whether it is reasonable for people to trust the claims of politicians and journalists based upon the evidence that is available to us. Where we find mistakes, we ask for them to be corrected.

Feedback:

We aim for our factchecks to be as accurate and up-to-date as possible. If you think we have made an error or missed some relevant information, please email: feedback@fullfact.org

Full Fact

Main menu

Can we trust anti-social behaviour statistics?

Liked this? Read these:

Can't get enough of Full Fact? Sign up to our monthly newsletter

Feedback:

More on these topics:

Search the site