Why we still shouldn’t trust the crime statistics
The notion of a measurable “real” crime rate is a myth. Crime is the product of the actions of, at least, four parties
The Central Statistics Office has started publishing crime statistics again this week. These had been suspended over uncertainties about the reliability of the figures, a concern that reached its zenith over the issue of homicide and the manner in which apparent homicides had been misclassified. The new data comes with a health warning describing the data as “under reservation”.
Despite the CSO is presenting this as a new problem the issues around the accuracy of crime statistics are as old as the notion of crime statistics themselves and not really resolvable.
The current focus of attention in Ireland has been on the period from 2003 onwards but the problem of unreliability of crime statistics has a longer but largely unacknowledged history.
There has always been a problem of the non-reporting of crime. This is particularly pronounced in the areas of sexual crime, where there has been an unwillingness to report and to classify child sexual abuse as a crime and similarly with rape where as little as 30 per cent of cases of sexual crime are reported to the gardaí.
There has also always been a problem of how crimes were recorded by the gardaí. This was made evident in a seminal study carried out by Richard Breen and David Rottman, then researchers in the ESRI, in the early 1980s. They asked a national sample of just under 9,000 individuals if they had been victims any of six criminal offences, including burglary, car theft, larceny from a vehicle and vandalism, offences that at the time accounted for over half the recorded crime in Ireland. The results of the survey suggested that the number of domestic burglaries was in region of 35,000. Of these 88 per cent of victims said they reported the offence to the gardaí, meaning that 31,000 were reported to them. The gardaí only recorded 17,000, a staggering difference. The problem lay in the failure of the gardaí to record all of the reported offences, an explanation that they strenuously rejected at the time and the research was largely ignored by policy makers and by the gardaí.
However these problems are not specifically Irish, though the scale of them might be. The quality of crime statistics in England and Wales has, for example, been an ongoing issue for the UK Statistics Authority. The problems are inherent in the nature of crime and in the means through which we record it. Having national crime figures is a reassuring piece of scientism, appearing to offering solid, incontrovertible and seemingly objective measures of the level of crime. There may be many in the world for whom statistics offer an aura of comforting certainty, a reliable yardstick against which to measure the competing claims of experts, public authorities and competing social interests but crime statistics are not one of these.
The notion of a measurable “real” crime rate is a myth. Crime is the product of the actions of, at least, four parties. The first is the state, which decides that certain actions are criminal, the second is people who engage in such behaviour, the third is the victims who report such offences, and the fourth is the police who investigate crimes. The actions of all of these are in principle measurable but in practice they are not. We can find out what kinds of actions are prohibited by the state, we can find out the number of crimes and the number of criminals who are identified, apprehended and processed by the police and the criminal justice system and we can find out the number of victims who report these offences to the police. But none of these is an accurate measurement of the “real’ level of crime.
No matter how crimes statistics are compiled they will always be unreliable, containing as they do, at least two different and irreconcilable kinds of data. One is that supplied, generally incompletely, by victims such as the number of violent crimes, property crimes, and thefts. The other is the number of recorded crimes that are a reflection of the level of police activity rather than the level of occurrence of such offences. This includes drugs offences, public order offences, road traffic offences or what the Irish Crime Classification system calls “dangerous or negligent acts”. The number of offences detected here is a reflection of the level of active enforcement rather than the level of occurrence of such offences.
So, for example, an increased priority given by the gardaí to drugs offences will paradoxically increase the number of such offences that are recorded. Additionally the mistrust caused among drugs criminals by the apparent success of police activities against them can lead to increased violence within the drugs gangs as mutual suspicion flares as to the possibility of there being police informers or compromised fellow criminals in their midst. This means that the relationship between police activity and the level of crime is a complex and by no means a uni-dimensional one.
The overall implications of these considerations are that it is not just current crime figures that must be treated with “reservations”, crime statistics should always be treated in this way.
Ciaran McCullagh is Adjunct Professor in the School of Law in the University of Limerick.