Serializing to reveal a norm of institutional behavior: a cross-case analysis of three jurisdictions.

Summary This article proposes a methodological reflection on the use of serialized data, based on a critical review of three research studies, conducted in history and economics, on judicial practices. These three studies differ in their disciplinary anchorage, but also in the period considered (from the Middle Ages to the present day) and in the jurisdiction studied (the royal court of Chatelet, the disciplinary councils of the advocate's office and the family court judges). However, they share the same motivation: to try to reveal, by the serialization of individual data, the norm of behavior of a judicial institution. This same research motivation makes it possible to compare the methodological choices made. It confronts the researcher, even for the contemporary period, with non-exhaustive information that he must construct himself from the archival material that he examines. The type of loss of information with which the researcher is confronted - which refers in part to the properties of the institutional context of the phenomenon studied - explains the methodological choices between the researches, independently of their disciplinary anchoring. Only the inferential use of the statistical method, i.e. the articulation between models and data, seems to reveal differences in methodological habit, without these differences covering disciplinary affiliations.