This course was shaped by the politically innovative, intellectually creative ferment at the Institut national de la statistique et des études économiques during the ten years following May 1968. The most innovative young administrators, with Desrosières as a driving force, refused to treat numerical data as a neutral "raw material" available for different uses. They believed that making such data available made sense only when accompanied by reflection on the social and technical conditions under which it was produced.
From then on, the production of numerical data and its uses, particularly by the state, lay at the heart of his research. He initially focused on how classification systems shape statistical information and, through it, the structure of society (see Les Catégories socioprofessionnelles (Socio-occupational Categories), co-authored with Laurent Thévenot and published by Éditions La Découverte in 1988). This line of inquiry drew on research that bridged sociology and statistics and led him to collaborate with Pierre Bourdieu.
His analysis of the role played by these categorization processes grew into an ambitious intellectual undertaking aimed at tracing the co-production of statistics, the state and the economy since the 17th century, drawing on both political history and the history of science. An initial synthesis of this research was published in 1993 as La Politique des grands nombres (The Politics of Large Numbers; La Découverte), a book that was subsequently translated into many languages.
His work on statistics as tools of knowledge and governance then turned to two issues of global importance. The first was a shift in statistics and indicators that involved taking account, in the production of numerical data itself, of the incentives created by their publication for those affected (benchmarking). The second concerned the social and political effects of new governance techniques based on so-called "randomized" experiments and claiming to identify the causal factors on which "decision-makers" should focus exclusively, particularly in poverty policy. These later studies were published in the two volumes of L’Argument statistique (The Statistical Argument), published in 2008 by Presses des Mines.
References:
• Prouver et gouverner: une analyse politique des statistiques publiques (Proving and Governing: A Political Analysis of Official Statistics). Alain Desrosières, La Découverte, 2014.
• Mathématiques et géographie (Mathematics and Geography). Bibliothèque Tangente 40, 2011.
