Approximations et erreurs dans les tables de nuptialité de générations

Henry Louis. ? Approximations and Errors in Generation Nuptiality Tables. Cohort analysis ? which was hardly used before the war ? is now prefered to period analysis for studying fertility and nuptiality; on the reverse, period analysis is still the more widely used, as it has not the same drawbacks as when used for fertility and nuptiality and as it gives a better answer to the problems. It is why more generation nuptiality tables should be found than generation mortality tables. Actually, as far as we can ascertain, there is no generation nuptiality table ?strictly speaking ? whereas there are generation mortality tables. As for nuptiality, only a few current tables and a few approximations of fundamental compendrum of generation table indexes are available. This state of affairs, slightly akin to a paradox when we come to think of it, is made more understandable by the rather minor importance of nuptiality in demographic studies and by the difficulties encountered when a nuptiality generation table is drawn according to methods similar to the classical ones used for drawing current tables. We faced those very difficulties when we started a study ? still under way ? on the perturbations brought into the nuptiality of French women ? born form 1881 up to 1910 ? by the 1914-1918 war. Those difficulties drew our attention to the approxi- mations already in use or which could be used ? as there is no table ? and on the effects of errors made during the observation. Quite soon we realised that it was worth while to write a paper on this subject, including the following headings : 1. A systematic report : a. On the aim of nuptiality analysis and its methods; b. On observation errors. 2. An illustration of those theories : a part of the results already obtained in the study of female generations born in France from 1881 up to 1910.