Review Society

The Numbers of Crime

About: Ph. Robert et R. Zauberman, Mesurer la délinquance, Presses de Sciences Po

When it comes to delinquency, the press, as well as political discourse, tends to emphasize the “crime angle” over the “sociological angle.” Moreover, when quantitative data is used in public debate, trust in statistics is low. Philippe Robert’s and Renée Zauberman’s new book may suggest a way out of this unfortunate situation.

Reviewed: Philippe Robert Renée Zauberman, Mesurer la délinquance, Paris, Presses deSciences Po, 2011. 177 p., 15 €.

When it comes to delinquency, the press, like political discourse, emphasizes the “crime angle” more than the “sociological angle.” Moreover, when quantitative data is used in public debate, trust in statistics is low. Philippe Robert’s and Renée Zauberman’s new book, published in Science Po’s La bibliothèque du citoyen (“The Citizen’s Library”) collection, may suggest a way out of this unfortunate situation. The fruit of more than a quarter of a century of research on delinquency at the Centre de Recherches Sociologiques sur le Droit et les Institutions Pénales (the Center for Sociological Research on Law and Penal Institutions, or CESDIP), this clearly-written book is aimed at a broad reading public. The authors offer a stern critique of the methodological weaknesses of French institutions that generate and disseminate data on delinquency, while at the same time making a plea for liberating this data from its dependence on penal statistics through a more systematic comparison of a variety of sources.

Constructivist Statistical Sociology

The authors take an original position: the goal is not only to discuss the reliability of crime data instruments, but to propose measures that account for the constructed character of delinquency and of the dependence of measures on a long chain of recordings. Though present throughout the book, this argument is stated most clearly in the first two chapters, where the authors explain the transition from a monopoly of penal statistics to the consideration of an array of sources.

In the eighteenth century, across Europe, the first measurements of delinquency were based on judicial statistics generated by counting court judgments. In France, the Compte général de l’administration de la justice criminelle (General Count of the Administration of Criminal Justice) was the primary source of crime statistics until the early 1980s, when it was definitively replaced by police data, “which quickly monopolized delinquency measurement” (p. 23). Yet whether crime is measured through juridical or police statistics, the same question begs to be asked: can one base measures of delinquency on administrative statistics that are primarily aimed at counting judicial and police actions? This question was raised as early as the first experiments with crime data gathering in the eighteenth century. They continue to inform the major epistemological controversies relating to crime statistics in our own day. The question draws its force from the “black number” of delinquency, which refers to the gap between known acts of delinquency and acts of delinquency actually committed. Yet as the authors point out: “the ‘black number’ has merely served as a verbal fig leaf for an unquestioned use of the penal figures to measure for delinquency” (p. 20).

The positivist metrology of delinquency could not withstand the numerous criticisms it encountered during the 1960s in the United States as well as Europe. The authors sum them up as follows: “It is impossible to measure delinquency in itself, but only enumerations that count, at a particular point in the process, the categorizations made by a range of lay or professional actors, who deemed certain forms of behavior to be criminal” (p. 58). From this principle arises a measuring practice based on a variety of sources, which notably includes seeking data in other sectors (health statistics, obligatory withholdings, etc.), learning from economics (a calculation of a crime in terms of money-equivalents), and surveys (victimization surveys and self-reported delinquency). The comparison of these points of view is crucial to the accurate measurement of crime.

Accepting Uncertainty through Comparison

Delinquency is not a homogeneous phenomenon. This is why the authors feel compelled to distinguish between two situations: delinquency without direct victims (illegal immigration, tax fraud, etc.) and delinquency with direct victims (theft, assault, etc.). In the first instance, comparison is particularly difficult, as penal institutions have no interest in seeking alternative measures in circumstances where they “have every interest in insisting on their statistics as ‘the’ measure of delinquency” (p. 69). The authors cite two examples to illustrate this situation: the misuse of public finances and the consumption of prohibited goods. Let us consider the latter case, in which the principle of confrontation is particularly limited as available sources (police statistics, Escapad investigations during the appeals day for preparing the defense of sixteen to seventeen year olds, and the Baromêtre Santé sources for eighteen to forty-four year olds) do not agree with one another: No immediate comparison is possible between police data […] and the results of investigations: the first provides the number of arrests, the other two trends within particular age brackets. We should nevertheless point out that these developments are not always similar: a major and sustained increased of police figures, an increase followed by a decrease of regular usage of cannabis among young people, or an increase followed by a plateau of current cannabis usage by eighteen to forty-four year olds. Police statistics accurately reflect the increased interest in this crime, but poorly reflect the actual evolution of cannabis usage. If one combines the lessons of these different investigations, they seem to indicate an increase in the final decade of the twentieth century, followed by a decrease (among the young) and stabilization (for adults) (pp. 72-73).

The authors devote a much longer analysis to delinquency with victims. This is an approach with which they are very familiar, involving the comparison of police statistics and victimization surveys. The authors focus on assault and property crime. Let us concentrate on the former (homicide, assault, violent theft, etc.), where the authors’ constructivist approach is particularly well illustrated, notably in the case of “aggravated assault” (i.e., grievous bodily harm). This category derives from a legal convention that has been in constant flux since the mid-nineties. Parliament has long considered as a crime any assault resulting in a loss of more than eight working days. Anything less was simply subject to a fine. Yet a series of laws gradually included violence that resulted in no work loss, in an attempt to suppress assault directed against specific professions (teachers, public transportation employees, etc.). In this context, “only victimization surveys are useful and only for measuring physical assault (as police statistics only measure legal inflation); for assault without grievous bodily harm, victimization surveys are not sufficiently trustworthy, as this category has never ceased to change.” Ultimately, the comparative approach does not so much indicate the actual state of delinquency as encourage us to agree on a statement, keeping in mind that “police data [which remain the main information source used in public debate] seem to overestimate the seriousness of serious crime and to underestimate low-intensity violence” (p. 92).

This kind of benchmarking can be used not only for comparing difference data sources, but also for comparisons across territories at various levels. Yet whatever its form, benchmarking is always used to shed light on the penal system’s input. Thanks to their constructivism, the authors manage with great skill to accommodate themselves to the inherent uncertainty of statistical data. Comparison requires that there be a “conversation” (i.e., a public debate) about every data collection. A public policy for measuring crime has, however, yet to be implemented in France: unlike in Great Britain, for example, France’s police administration continues to maintain under a veil of secrecy the way in which data is recorded and cares little about ensuring the continuity of victimization surveys.

The Production and Use of Measures

Constructivism and the many limits of the comparisons analyzed throughout the book do not prevent the authors from offering, in their final chapter (“From Measures to Diagnostics”), a general synthesis on the state of delinquency in France. They do so by being particularly attentive to the overall coherence of the comparative picture. Using the comparison of sources to identify broad trends proves to be a daunting challenge. For major trends are diagnosed: 1/ a powerful wave of property delinquency (1960-195), corresponding to the shift towards a society of mass consumption which becomes, at the same time, a society of mass predation—a wave that has retreated over the past two decades; 2/ two kinds of more recent violence: forcible theft, corresponding to a displacement effect due to targets becoming more difficult with the diffusion of situational preventative procedures; and acts of low-intensity expressive aggression, which is incomprehensible outside of marginalized spaces, cut off from social networks, where “one sees the rise of a problematic of reputation very close to that of honor as it prevails in traditional societies. To defend one’s reputation against marginalization’s brutal wounds, one can rely on no other resources than one’s own body, through verbal virtuosity or physical force”; 3/ an impressive rise of acts of public delinquency: road crime, attacks on police officers, illegal immigration, and the production and use of prohibited goods. One can understand this explosion only if one recognizes that it takes on an increasingly important place in acts of delinquency recorded by the police, in a context of greater state concern; 4/ and, finally, the scarcity of information of economic and financial delinquency.

If the interest of this diagnosis lies in the way it calls into question the prevailing orthodoxies purporting to explain delinquency, it is nonetheless limited by the fact that authors have not submitted their own diagnosis to the “conversation requirement.” The measure/diagnosis distinction implies that that there is statistical knowledge on the one hand and its application on the other. Yet each point that the authors diagnose is the object of a debate or scientific controversy, in which actors employ statistical languages other than quantification (a notion that is more accurate than that of the measure; see the works of Alain Desrosières on this point). Statistical modeling, for instance, makes it possible to nuance the explanations of mass predation in many ways. One might mention in particular the work of Tim Hope, which drew on databases of British victimization surveys: by using models inspired by econometrics, Hope tries to explain property crime in terms of the victims’ practical “knowledge,” i.e., their ability to resolve unpredictable problems and to find contextual solutions. [1] Moreover, the recent use in the realm of developmental criminology of randomized control trials and longitudinal studies on groups of delinquents offers another explanation of more recent violence: persistent aggressive behavior in a minority of children is a predictor of chronic adolescent delinquency. This type of “measure” of violence is indistinguishable from explanations that deny sociological factors, in favor of personal traits and family environment. [2] In short, in presenting the numerous controversies surrounding the cause of delinquency, the authors might have extended their pragmatic view of measures to the diagnosis itself.

Yet these remarks should in no way detract from this work’s scholarly as well as political merits. Without popularizing or oversimplifying the issues, it seeks to make a lay public interested in the preoccupations of statisticians. The book is thus an excellent tool for democratizing delinquency measures.

First published in Translated from French by Michael C. Behrent with the support of the Institut français.

by Bilel Benbouzid, 8 May 2012

To quote this article :

Bilel Benbouzid, « The Numbers of Crime », Books and Ideas , 8 May 2012. ISSN : 2105-3030. URL :

Nota Bene:

If you want to discuss this essay further, you can send a proposal to the editorial team (redaction at We will get back to you as soon as possible.


[1Hope, T. and Trickett, A. (2004), “La distribution de la victimation dans la population,” Déviances et Société 3:28, 385-404.

[2Tremblay, R.E.(2003). “Why Socialization Fails? The Case of Chronic Physical Agression.” Lahey, B.B.; Moffitt,A. Caspi, A. (eds.), The Causes of Conduct Disorder and Juvenile Deliquency, Guilford Press.