Dyscalculia is a term that denotes difficulty in processing numerical information, particularly mathematical operations. It can be acquired (e.g., in the case of persons who knew how to perform calculations but who largely lost the ability to do so, for instance after a neurological accident), or developmental, if it involves persons incapable of learning to calculate at a normal level. While developmental dyscalculia may affect both children and adults, the acquired form should be almost exclusively confined to adults. As it happens, there are hardly any studies of the occurrence of dyscalculia among adults in France, whereas there are many studies on children and adolescents. This article therefore attempts to estimate the proportion of adults whose performance is consistent with dyscalculia, using results of the 2004 Information and Daily Life Survey (Information et Vie Quotidienne: IVQ). This identification combines a criterion of calculation difficulty and a criterion of divergence between performance in mathematics and French tests of comparable level. We do assume a “singularity” of poor calculation skills, as distinct from a common weakness in both of these instrumental forms of knowledge. However, we can speak only of persons “potentially” affected by dyscalculia. The 2004 IVQ Survey allows no corroboration against other tests, such as neurological tests. Other hypotheses besides “dyscalculia” may be put forward, starting with those pertaining to the administration of the survey: respondents may be tired at the end of the interview, when they are invited to perform the calculation items. Our test identifies 387 respondents who fulfil the two criteria, out of the 10,213 in the 2004 IVQ sample that we were able to analyze. From these figures, we estimate the proportion of French adults potentially affected by dyscalculia at 2.95%, of whom 0.55% suffer from a “deep” form of the disability.