Analyse factorielle et construction des variables. L'origine géographique des auteurs anglais (1300-1600)

Abstract

In a prosopographical study, some variables may prove difficult to use when the number of their values is so large that it divides the population into small groups. It is therefore necessary to group them together. But how to do this when there are no obvious categories which make it possible to achieve such groupings? It is precisely what happens with the place of birth in England: the county is usually too small, and regions are lacking in historical legitimacy and stability. We suggest to construct new variables based upon the proximity of the counties profiles produced by a factorial analysis, crossing place of birth and time. The regions engendered by this process prove more coherent and easier to interpret than the traditional « geographical » regions, as two further factorial analysis show. © Éditions de l'EHESS.