The relevance of Bayes’ theorem in probability theory

Author: Roberts, P., Professor
Abstract: It is suggested that conventional statistical significance tests are inappropriate for research in astrology and that Bayesian methods are better. Bayes’ theorem is explained using probability tree notation and examples from everyday life. It is argued that Michel Gauquelin could have inadvertently biased his work on planets and character traits – but still have been right. Examples are offered from the history of science to show that prominent scientists have often embellished results to make them more convincing. Finally, a Bayesian analysis of the Astro-Questionnaire results taken in conjunction with Gauquelin’s character trait work indicates that a prior probability of 1/5 develops to a post probability of nearly 2/3. It is proposed that this analysis gives support to the contention that further experimental work on the character trait hypothesis is needed.
Keywords: Bayes, Gauquelin, questionnaire, trait
Publication: Correlation: Astrological Association research journal in astrology
Issue: Volume 20 Issue 2
Dated: 2002
Pages: Pages 49 – 57

Posted in Free Research Abstract