Saturday, November 22, 2008

The Relative Accuracy of Crime Stats

A recent report from the OND (National Observatory on Crime), summarized in Le Figaro, confirms the reliability of police statistics on theft, but discloses that four out of five acts of physical violence, though reported, do not lead to a formal complaint:

The OND conducted a titanic investigation, questioning victims in hundreds of interviews, resulting for the first time in two separate studies that finally allow for comparisons.

Fewer than 20% of acts of physical violence (domestic violence excepted) resulted in a formal complaint in 2007. Out of 1.4 million acts reported, only 273,00 were investigated.

Contradicting somewhat the declarations of the Interior Ministry last January, in which the "first decrease in violence in ten years" was noted, the OND affirms instead that the figures for 2007 are "higher than those for 2006." "Even though this variation is not considered as significant," they add cautiously.

The OND estimates at 200,000 the number of victims who have suffered repeated attacks in the course of the year. Taking only the last act inflicted, the OND notes that only four persons out of ten went to the police or gendarmes to report it. Out of these four, three filed a formal complaint, while the fourth filed only a report in the police registry for minor acts that do not go to court. When the six persons who did not go to the police at all were questioned, 39% responded that they did not go "because it wasn't serious", and 28% "because it would have served no purpose."

Another piece of evidence flushed out by the OND: the discrepancy in the number of burglaries formally reported by victims - 240,000 in 2007, and the official figures - 150,000. One hypothesis is that the authorities initially made errors in the classification of these crimes.

The good news is that theft continues to decrease in France, by more than 7% in one year. This is due mainly to the "decrease in auto theft", says Christophe Ouliez, a leading analyst at the OND.

The OND wants to be fair to the Interior Ministry. Its estimate of the number of formal complaints filed last year for theft and attempted theft corroborates the official figures: 1,644,000 according to criminologists, vs 1,651,259 according to the police. On this point, at least, the long-suspected inaccuracy of the ministry's reports is unfounded.

There is a companion article to this one that describes the typical victims. Here are excerpts:

The OND report establishes a direct link between unemployment and the risk of violence. Almost four unemployed women out of a hundred had suffered at least one act of violence inflicted on them by their partner. Statistician Cyril Rizik insists that the figure is important and that the danger increases if the partner is also unemployed.

Men and women are equally victims, but women are primarily victimized by their partner or ex-partner. Men between 18 and 25 are attacked primarily by unknown aggressors.

The OND refutes a certain number of prejudices. Persons living in the projects are not more violent than other tenants. But the residents of the ZUS (Sensitive Urban Zones) are, despite everything, twice as exposed as couples in rural zones. Women with a certificate of basic studies are more likely to be victims of their partner than those with more advanced diplomas.

However the interesting thing concerns men who are victims of their partner:

Women living with a partner are more likely to commit acts of violence when they (the women) have diplomas. The risk for a man is three times higher if the woman is a college grad, and twice as high if she has a high school diploma.

I suppose this implies that better educated women have higher expectations and lower levels of tolerance than the less well educated. It may imply too that the female college grad is more of a feminist and more likely to say to herself that whatever he can do she can do also. Simultaneously, she must harbor the belief that he will not fight back. None of this means that she is more realistic than her under-educated "sisters". Education often has the effect of raising expectations beyond a reasonable level, and thus triggering shattering disappointments.

Remember, the above is speculation...

Can we trust the OND? A visit to its website reveals that it is a department of the INHES (National Institute of Advanced Studies on Security), which in turn is part of the Interior Ministry. It is defined thus:

The National Observatory on Crime (OND) is a department of the National Institute of Advanced Studies on Security (INHES), a public establishment, administrative in nature. Its administration is placed directly under the authority of the director of INHES, assisted by a department head. Its strategic independence is assured by an orientation council.

The last line above is the critical point - it must be independent to be credible. Since it has disputed some of the crucial figures issued by the Interior Ministry, it can be assumed that it is, at least to some extent, impartial.

A reminder that, for the French, "administration" refers to the government.

Figures on crime, like figures on immigration, are just about impossible to verify accurately. And the one thing statisticians cannot report on is the atmosphere of crime, the fear and apprehension in which people live, even when no crime is committed. You are always more likely to emerge from a subway ride alive, but the inner knowledge that a crime could happen at anytime causes stress and discouragement in the everyday lives of the citizens, not to mention what they have to look at, listen to, and put up with in their neighborhoods, schools, offices, buses, etc... Crime is more than a knifing.

For those interested (and with time on their hands), there are two reports (pdf format) in English available at the INHES website. One is a 33-page report on crimes and petty crimes in 2006-2007. The second is a 3-page report focusing on acts of physical violence committed in Paris between March and May 2006. This was the agitated period following the riots of the previous year.

Labels: , ,


Post a Comment

Links to this post:

Create a Link

<< Home