My suspicion that government crime statistics could be misleading (see previous entry) is supported by a study published by Civitas in June last year.

It appears that the British Crime Survey has a very particular way of counting crimes. The real incidence of all violent crime appears to be 83 per cent higher than that which given in the British Crime Survey. This understating of crime has been going on since the survey started in 1981. Since the total level of crime in each year since then has been understated but to an unknown but presumably varying degree each year, the assertion that violent crime is going down is not wholly reliable. I suspect there are further reasons to doubt the trustworthiness of the crime figures. A few have already been suggested in comments on my previous post (below).

I should add that the academics who wrote this report went out of their way not to criticise the statisticians themselves. The fault they find is with a way of treating the figures that was started in 1981.

Here is part of the Civitas press release:

...ever since its inception in 1981, the British Crime Survey (BCS) has omitted many crimes committed against people who have been repeat victims. If people are victimised in the same way by the same perpetrators more than five times in a year, the number of crimes is put down as five. The justification for this was ‘to avoid extreme cases distorting the rates’, but, as Farrell and Pease point out, ‘if the people who say they suffered ten incidents really did, it is capping the series at five that distorts the rate’.

By recalculating the figures without the arbitrary cap of five crimes, Farrell and Pease have revealed that there are over three million crimes omitted from the BCS:

In its most recent published sweep, BCS estimated an annual total of some 6.8 million ‘household’ crimes (covering burglary; theft in a dwelling; other household theft; thefts of and from vehicles; bicycle theft; and vandalism to household property and vehicles). It estimated some 4.1 million ‘personal’ crimes (which covers assault, sexual offences, robbery, theft from the person, and other personal theft). Our re-analysis reveals that, if we believe what the respondents tell us, there would be 7.8 million household offences and 6.3 million personal crimes. Thus, removing the arbitrary five offence limit, over three million extra offences come to light… Household crime is increased by 15% and personal crime by a staggering 52%. As the sum of personal and household crimes, total crime would have been understated by 29%.

The increase in the number of crimes is not evenly spread across all types of crime. For example, theft of vehicles is not increased at all, but levels of vandalism are almost a quarter more than reported, and there are 20 per cent more burglaries. Violent crime of all types increases by 83 per cent. Violence perpetrated by an acquaintance increases by 156 per cent and domestic violence by 140 per cent. As Farrell and Pease say, ‘these are not minor differences’.

The full press release is here.

Posted by James Bartholomew • Indexed in Behaviour & Crime

Comments (0) TrackBack (0)