Early Trace of the Current Hypothesis Test: Graunt (1662), Arbuthnott (1710) and Sagravesande (1712)

Cristina Camúñez Díaz, Consuelo Borrero Gallego, Ana Soltero Díaz  
(Universidad de Sevilla, Spain)

Abstract: The hypothesis test, as today we understand, are born in the first decades of the twentieth century, with the work of Fisher, on the one hand, and Neyman and Pearson, on the other, which were developed from antagonistic philosophical positions, which they have been maintained over time. But for two centuries before there were attempts to endorse assumptions made from reality with the available data. Thus, from the accounts mortality weekly published the City of London, John Graunt (1662) he believed prove that “the unhealthiest years are less fertile” (exists in time a negative correlation between the annual number of burials and the birth), while John Arbuthnott (paper read in 1711 and published in 1712), gave “Arguments for Divine Providence” and sGravesande a mathematical proof that “God directs what happens in the world”, in both cases, from the regular rate of births of men and women. Although the theoretical foundations of these works are of dubious rigor, proposals are ingenious and nearby, in some respects, what we do today when we perform a hypothesis test. This paper analyzes in detail the early contributions of these three authors by looking at what it looks like now and pointing out the mistakes made by them.

Key words: Graunt, Arbuthnott, sGravesande, hypothesis test, regularity of births among women men

JEL codes: C000, C120