Download PDF
Homeopathy 2008; 97(04): 167-168
DOI: 10.1016/j.homp.2008.09.003
Editorial
Copyright © The Faculty of Homeopathy 2008

Bayes, bias and plausibility

Peter Fisher
Subject Editor:
Further Information

Publication History

Publication Date:
16 December 2017 (online)

Also available at
    Permissions and Reprints

Thomas Bayes was an 18th century English clergyman and mathematician, his most influential work ‘Essay Towards Solving a Problem in the Doctrine of Chances’ was published posthumously, and his ideas became influential only some two centuries after his death. Bayes described ‘conditional probabilities’ in terms of beliefs and degrees of uncertainty, his theorem is now mostly commonly used to calculate the influence of observation on so-called ‘prior’ probabilities, the probability after taking the new observation into account is termed ‘posterior’ probability. In this issue of homeopathy, Rutten cites Woodworth's example showing how the introduction of successive pieces of evidence might have influenced the jury's view of the probability that OJ Simpson was guilty of murder. The point is that jurors' subjective beliefs and attitudes (in this case to the police and their evidence) are decisive.[ 1 ]

Lex Rutten and his colleagues in the Committee for Methods and Validation of the VHAN (Dutch Homeopathic Physicians' Association) specialise in the application of Bayes' Theorem to homeopathy. They have shown how, in theory, this method can be used to revise and make more reliable the homeopathic repertory. They pointed out that the grading of indications should be based on their frequency relative to the general population, not on absolute occurrence using of absolute occurrence could mean that, for commonly used medicines, symptom listing in a rubric could actually be contraindications to the medicine in question! Introducing likelihood ratios obviates this problem and its converse: underestimating the significance of ‘peculiar and characteristic’ symptoms.[ 2 ] More recently they have demonstrated the method empirically, focussing on the symptom ‘sensitive to injustice’, found in 10% of over 3000 patients studied.[ 3 ]