45

Until the 1970s, studying access opportunities to higher education – and their historic evolution –, privileged measures based in the comparison of statistical deviations. This “orthodoxy” was succeeded by another, which followed difference quotients. Finally, from the 1980s onward, a “logistical orthodoxy” relying on odds ratios became common. This article critically analyzes the statistical treatment of fortune, by re-situating the competitive issues between various statistical methods. The article shows, through the study of a single example, that different statistical methods can lead to divergent conclusions; and that the succession of methodological orthodoxies, seemingly governed by technical considerations, were not immune to changes in the national and international politics of education. The ensemble of this analysis brings a critical and reflective regard to the increasing autonomy of a statistical sociology organized around abstract and sophisticated practices over the past thirty years.