Conclusion

Abstract

Within the limits imposed on monographs in this series, the author has contented himself with a summary of recent progress in demographic analysis. The classical sources, which in any event are cited in all collective works in quantitative demography, are left aside to avoid repetition of what is sufficiently well expounded elsewhere. To complete the plan of the present monograph it remains only to indicate a device to which the author has alluded in the first pages. We had remarked there that graphic representations, while rendering good service, are insufficient for expressing the complicated relations of demographic analysis, not only because of the large number of variables that enter into play, but even more because of the form of these relations. To effectively represent probabilistic relations it is not immobile graphics inscribed on paper that we need. We should rather construct mobile models which by their operation produce a collection of characteristics having a distribution of frequencies resembling the characteristics of the population to which they apply. Our ideal would be a collection of urns giving, by a draft effected in accord with prescribed rules, a true sampling of the characteristics of the population.^* That ideal is unrealizable in every detail, but it can be approached to a degree, depending on the one hand on our statistical resources, and on the other on the effort we are ready to give to the job.