Comment tenir compte des erreurs de mesure dans l’estimation de la stature des conscrits français ?

Scrutinizing the French conscription reports by department reveals that in a great many cases the distribution of draftees’height was abnormal to the point of affecting the meaning of results as indicators of “biological living standards” used in anthropometric research. The present article traces the sometimes-absurd distribution to the proportion of absentees reported at the examination for the class of 1906. Alternatively, such discrepancies can be attributed to reporting errors – either intentional or not. They illustrate that military service obligations were far from being uniformly accepted and enforced across the country and that the tolerance – or powerlessness – of authorities towards draft-dodging varied across departments. As a consequence scholars should exert great care in the handling of such indicators in the case of obviously biased samples and prefer for those apparently suitable to rely on the median rather than the average height.