Large-scale significance testing of the full Moon effect on deliveriesReportar como inadecuado

Large-scale significance testing of the full Moon effect on deliveries - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

1 Department of Economics, Ecole Polytechnique 2 CREM - Centre de Recherche en Economie et Management 3 IRMAR - Institut de Recherche Mathématique de Rennes 4 AGROCAMPUS OUEST

Abstract : As mentioned by many authors, the belief that the number of women going into labor and giving birth is higher during the full Moon is widespread, even among the medical staff. However, various statistical studies of the daily number of births along the Moon cycle, mostly on rather short periods from 40 to 60 lunar cycles, i.e. less than 5 years, conclude to contradictory results, which strengthens the need for a powerful analysis on a large amount of data. We propose a large-scale signficance testing analysis of the full Moon effect in each lunar cycles from 1968 until 2005 based on the daily numbers of births in France. A multiple testing methodology which accounts for dependence among lunar cycles is used to guarantee both a high overall power and a control of the False Discovery Rate at a low level. Results confirm the existence of a small yet marked full Moon effect: on average, one cycle per year shows a significantly larger birth rate during a 6-days period around the Full moon day than the other days of the cycle, which is four times more than for a comparison between any other 6-days period and the rest of the cycle.

Keywords : Number of births Full Moon effect High-dimensional data Multiple testing

Autor: Arthur Charpentier - David Causeur -



Documentos relacionados