Political Arithmetic Becomes a Discipline of Its Own

Statistics is the discipline of collecting, analyzing and interpreting numerical data well enough to distinguish a genuine underlying pattern from one that random variation alone could produce. Its roots lie in the same 1662 work that founded demography, John Graunt's analysis of London's mortality bills, an approach contemporaries called political arithmetic, and the field grew across the following two centuries into a formal mathematical discipline with its own theory of probability, sampling and inference, distinct from the practical population counting that had first prompted it. Pierre Simon Laplace and Carl Friedrich Gauss gave that theory of probability its rigorous mathematical footing in the early nineteenth century, and Ronald Fisher's work on experimental design and significance testing in the 1920s gave statistics much of the working method it still uses today.

Virtually every quantitative claim the atlas encounters about religion, how many adherents a tradition has, whether a practice correlates with a measurable outcome, whether a survey result generalizes beyond the people actually surveyed, is a statistical claim whose strength depends on sample size, sampling method and the size of the effect relative to its margin of error. Statistics is the discipline the atlas leans on to state plainly when a widely repeated number or correlation is well supported and when it is not, rather than repeating a figure because it is frequently cited.